NEURONAL BC1 RNA: CO-EXPRESSION WITH GROWTH-ASSOCIATED

PROTEIN-43 MESSENGER RNA

Y. LIN,a J. BROSIUSc and H. TIEDGEa,b*aDepartment of Physiology and Pharmacology, State University of New York, Health Science Center at Brooklyn, Brooklyn,

New York 11203, USAbDepartment of Neurology, State University of New York, Health Science Center at Brooklyn, Brooklyn, New York 11203, USA

cInstitute of Experimental Pathology/Molecular Neurobiology, University of MuÈnster, D-48149 MuÈnster, Germany

AbstractÐBrain-speci®c cytoplasmic RNA 1 (BC1-RNA), a non-coding RNA polymerase III transcript, is a neuronalRNA that is speci®cally targeted to dendritic domains. It is co-localized with components of the dendritic protein syntheticmachinery, and it has been suggested to operate in the regulation of local translation-related processes in postsynapticmicrodomains, thus subserving long-term synaptic plasticity in neurons. To probe the relevance of BC1 expression inneuronal plasticity, we have analyzed the expression pattern of BC1 RNA in the rat nervous system. We found that BC1RNA is expressed by a speci®c subset of neurons (but not by non-neuronal cells) in the central and peripheral nervous systemof the adult rat. The BC1 labeling pattern indicates that the subcellular location of the RNA is typically postsynaptic which,depending on cell type, manifests itself in a predominantly somatic, somatodendritic, or dendritic location. Our resultsfurther show that BC1-expressing neurons typically co-express the messenger RNA for growth-associated protein-43 (GAP-43). Such co-expression was observed in diverse brain areas, including the olfactory bulb, neocortex, and hippocampus,among others. While BC1 RNA was in many neuronal cell types detectable in distal dendritic domains, GAP-43 messengerRNA was typically more restricted to neuronal perikarya.

In the mature nervous system, expression of GAP-43 has been described as an intrinsic determinant of predominantlypresynaptic plasticity, while BC1 RNA has been implicated in postsynaptic plasticity. Co-expression of both RNAs, asreported here, thus identi®es a distinct subset of neurons in the rat nervous system that exhibits both types of plasticity.q 2001 IBRO. Published by Elsevier Science Ltd. All rights reserved.

Key words: gene expression, RNA localization, dendrites, synapses.

The heterogeneity of synaptic connections along den-dritic arborizations, and the need for fast and independentregulation of such synapses, necessitate rapid and ef®-cient means to modulate local synaptic protein reper-toires. While a variety of mechanisms is likely to beinvolved in synaptic plasticity, dendritic protein synth-esis has been implicated in the long-term regulation oflocal postsynaptic functional architecture.1,12,18,29,32 Thishypothesis is now supported by several lines of evidence.(1) Polyribosomes are selectively located beneath post-synaptic membrane structures in dendritic spines.30 (2)Protein synthetic machinery, including components ofthe rough endoplasmic reticulum and the Golgi appara-tus, has been identi®ed in dendrites.10,33,36 (3) Varioustypes of RNA have been observed to be selectively local-ized to dendrites, and the expression and/or dendriticlocalization of some of these RNAs have been shownto be dependent upon neuronal activity.11,12,27,29,32

Dendritic RNAs include various mRNAs, e.g. onesthat encode cytoskeletal proteins, kinases, and receptors,among others.12,29,32 They also include RNA polymeraseI transcripts (rRNAs)13 and RNA polymerase III tran-scripts such as tRNAs33 and neuronal BC1 RNA.35

BC1 RNA is a non-coding activity-regulated RNA19

that is selectively and rapidly transported to dendritictarget sites.20 It has been detected at high levels inpreparations of postsynaptic dendritic microdomains.7,24

Brain-speci®c cytoplasmic RNA 1 (BC1 RNA) is co-localized with protein synthetic machinery in suchmicrodomains, and the combined evidence has led tothe hypothesis that the RNA, in the functional unit of aribonucleoprotein particle,6,14 plays a role in regulationof translation-related processes at postsynaptic sites.3

Dendritic mRNAs are heterogeneously distributed inrat brain as different mRNAs vary in terms of theirregional and cellular expression patterns, and in theirdendritic extents.22 To compare BC1 expression withthe distribution of these and other plasticity-relevantgene products, it was therefore necessary to establishthe expression pattern of BC1 RNA in the rat nervoussystem. Although the distribution of BC1 RNA doesnot apparently coincide with any previously describedneuronal cell type or transmitter-receptor system, our

BC1/GAP-43 expression 465

465

Neuroscience Vol. 103, No. 2, pp. 465±479, 2001q 2001 IBRO. Published by Elsevier Science Ltd

Printed in Great Britain. All rights reserved0306-4522/01 $20.00+0.00PII: S0306-4522(01)00003-3

Pergamon

www.elsevier.com/locate/neuroscience

*Corresponding author. Tel.: 11-718-270-1370; fax: 11-718-270-2223.

E-mail address: tiedge@hscbklyn.edu (H. Tiedge).Abbreviations: BC1 RNA, brain-speci®c cytoplasmic RNA 1;

CaMKII, calcium-calmodulin-dependent protein kinase II; GAP-43, growth-associated protein-43-kD; IP3, inositol 1,4,5-trisphos-phate; SSC, standard saline citrate.

initial data indicated a large degree of correspondencebetween the neural expression patterns of BC1 RNAand growth-associated protein-43 (GAP-43) mRNA.GAP-43, a phosphoprotein associated with the presynap-tic axonal membrane, has been shown to play a role inaxonal outgrowth in development and during regenera-tion.2,28 It has further been suggested that in the adultbrain, GAP-43 is expressed in areas that are associatedwith memory functions, and that it mediates experience-dependent neuronal plasticity in such areas.2 The proteinwas found concentrated in presynaptic axonal terminals2

whereas expression of GAP-43 mRNA is typicallyrestricted to neuronal somata and proximal-mostdendritic regions.4,16 In this paper, we examined thedistribution of BC1 RNA in the rodent nervous system,with particular attention given to patterns of co-expressionwith GAP-43 mRNA. Co-expression of BC1 RNA, a

postsynaptic RNA, and GAP-43, a predominantly pre-synaptic protein, de®nes a set of neurons that is charac-terized by high degrees of both pre- and postsynapticplasticity.

EXPERIMENTAL PROCEDURES

Procedures involving animals were in compliance with theNational Institutes of Health Guide for the Care and Use ofLaboratory Animals and were approved by the InstitutionalAnimal Care and Use Committee. All efforts were made to mini-mize animal suffering, and only the number of animals was usedthat was necessary to generate reliable data. Rats (Sprague±Dawley, male, about six-weeks old) were perfusion-®xed, and10-mm cryostat sections were collected on gelatin/poly-l-lysinecoated slides as described previously.31

In vitro transcribed RNA probes were 3H- or 35S-labeled.Probes were generated from linearized templates as described.31,35

Transcription vector pMK1 contains a sequence that represents

Y. Lin et al.466

Abbreviations used in ®gures

2n optic nerve3V 3rd ventricle4V 4th ventricle7n facial nerve or its rootI to VI layers I, II, III, IV, V and VI of the neocortexAA anterior amygdaloid areaaca anterior commissure, anterior partAq aqueduct (Sylvius)AV anteroventral thalamic nucleusbp brachium pontis (stem of middle cerebellar

peduncle)BSTM bed nucleus of the stria terminalis, medial divisionCA1 ®eld CA1 of hippocampusCA2 ®eld CA2 of hippocampusCA3 ®eld CA3 of hippocampusCb cerebellumcc corpus callosumCe central amygdaloid nucleusCGA central gray, alpha partCGB central gray, beta partchp choroid plexusCI caudal interstitial nucleus of the medial

longitudinal fasciculusCIC central nucleus of the inferior colliculusCLi caudal linear nucleus of the rapheCPu caudate putamen (striatum)cst commissural stria terminalisCx cerebral cortexDCIC dorsal cortex of the inferior colliculusDEn dorsal endopiriform nucleusDG dentate gyrusDpMe deep mesencephalic nucleusDR dorsal raphe nucleusec external capsulef fornix® ®mbria of the hippocampusfmj forceps major of the corpus callosumfr fasciculus retro¯exusgcc genu of the corpus callosumGrDG granular layer of the dentate gyrusHDB nucleus of the horizontal limb of the diagonal

bandic internal capsuleInfS infundibular stemLC locus coeruleusLD laterodorsal thalamic nucleuslfp longitudinal fasciculus of the ponsLGP lateral globus pallidusLH lateral hypothalamic area

LHb lateral habenular nucleusLSD lateral septal nucleus, dorsal partLSI lateral septal nucleus, intermediate partLSV lateral septal nucleus, ventral partLV lateral ventricleLuc stratum lucidum of the hippocampusmcp middle cerebellar peduncleMePV medial amygdaloid nucleus, posteroventral partMHb medial habenular nucleusMnR median raphe nucleusMol molecular layer of the dentate gyrusMPB medial parabrachial nucleusMPO medial preoptic nucleusmt mammillothalamic tractMVPO medioventral periolivary nucleusopt optic tractOr stratum oriens of the hippocampusox optic chiasmPAG periaqueductal grayPDTg posterodorsal tegmental nucleusPFl para¯occulusPH posterior hypothalamic areaPir piriform cortexPMCo posteromedial cortical amygdaloid nucleus (C3)Pn pontine nucleiPnO pontine reticular nucleus, oral partPnR pontine raphe nucleusPr5 principal sensory trigeminal nucleusPT paratenial thalamic nucleusPV paraventricular thalamic nucleusPVA paraventricular thalamic nucleus, anterior partPVP paraventricular thalamic nucleus, posterior partPy stratum pyramidale of the hippocampuspy pyramidal tractRad stratum radiatum of the hippocampusRt reticular thalamic nucleusS subiculums5 sensory root of the trigeminal nervescp superior cerebellar peduncle (brachium

conjunctivum)SFO subfornical organSO supraoptic nucleussp5 spinal trigeminal tractSuG subgeniculate nucleusVCA ventral cochlear nucleus, anterior partVLG ventral lateral geniculate nucleusVMH ventromedial hypothalamic nucleusZI zona incertaZo zonal layer of the superior colliculus

the 60 3 0-most nucleotides of BC1 RNA.35 GAP-43 mRNAprobes were transcribed from a previously described vector.21

Prior to hybridization, sections were UV-illuminated (wide-spectrum UV light, germicidal UV lamp, 30 W) at a distance of30 cm for 12 min.31 Sections to be hybridized with GAP-43mRNA probes were in some cases treated with proteinase K[Roche, Indianapolis, IN; 5 mg/ml in 2 £ standard saline citrate(SSC); 1 £ SSC� 150 mM NaCl, 15 mM sodium citrate pH 7.5)for 10 min at room temperature, followed by three rinses, 5 mineach, in 2 £ SSC at room temperature. However, since we wereunable to detect differences in signal intensities between protein-ase K treated and untreated sections, proteinase K treatment wasusually omitted. Proteinase K treatment is also not required forBC1 probes (probe length ,100 nucleotides). Prehybridization,hybridization, and post-hybridization treatments were carried outas described.31

Dried sections were exposed to Fuji RX X-ray ®lm for auto-radiography. For microscopy, sections were dipped in NTB2nuclear track emulsion (Eastman Kodak, Rochester, NY), diluted1:1 with high-performance liquid chromatography water.Sections were allowed to dry at 18±228C for 2 h, exposed at48C for three days (BC1 RNA probes, 35S-labeled), 21 days(BC1 RNA probes, 3H-labeled), or 18 days (GAP-43 mRNAprobes, 35S-labeled). Sections were counterstained with CresylViolet and inspected under dark-®eld optics (Nikon Microphot-FXA). Densities of autoradiographic silver grains were used as anindicator of signal intensities. Micrographs were taken withKodak Ektachrome 160T tungsten ®lm.

In all experiments, speci®city of BC1 and GAP-43 probes wasascertained by sense-strand control experiments performed inparallel. These controls never produced higher-than-backgroundsignals in any of the brain or other areas analyzed (data notshown).31,35

RESULTS

Expression pattern of BC1 RNA in the adult rat nervoussystem

We analyzed the distribution of BC1 RNA in the adultrat brain by in situ hybridization, using radiolabeledprobes speci®c for the 3 0 unique region of the RNA.The results are presented in the form of ®lm autoradio-graphs for overview and in the form of emulsion auto-radiographs for detail. Fig. 1 shows expression of BC1RNA in a series of coronal sections through the adult ratbrain. The distribution of the RNA follows a complexpattern that appears to be unrelated to any particularneuronal cell type. Prominent among the brain areaswith high BC1 labeling intensities was the neocortex.The BC1 signal appeared layered, re¯ecting the layeredstructure of the neocortex (see also below for emulsionautoradiography). On the other hand, variations in theautoradiographic signal between different neocorticalregions were less signi®cant (as is evidenced, forinstance, by a comparison of occipital and temporalcortex). Similarly strong BC1 labeling intensities wereobserved in various elements of the amygdaloidcomplex, including nuclei in the olfactory, medial,central, and basolateral amygdala, as well as the bednucleus of the stria terminalis. Intense labeling wasalso observed in the septal nuclei. In contrast, onlymoderate labeling was evident in the corpus striatum.Labeling was structured in layered areas (e.g. olfactorybulb, hippocampus, and cerebellum) some of which willbe discussed in more detail below.

A number of thalamic nuclei were strongly labeled,

among them the paraventricular thalamic nucleus, theparatenial thalamic nucleus, and the medial habenularnucleus of the epithalamus. A similarly strong hybridiza-tion signal was observed in several hypothalamic nuclei,including the supraoptic nucleus, the paraventricularhypothalamic nucleus, the dorso- and ventromedialhypothalamic nuclei, and several preoptic nuclei. In thevisual system, intense labeling was observed in theventral lateral geniculate nucleus (the dorsal lateral geni-culate nucleus was only moderately labeled), and in thesuperior colliculus, especially in the zonal and super®cialgray layers. Other strongly labeled midbrain areasinclude the inferior colliculus, in particular the dorsalcortex, and the central gray. BC1 labeling was moderatein the cerebellar cortex. In the brainstem, the nucleus ofthe solitary tract and the spinal trigeminal nucleus wereamong the most intensely labeled areas, but appreciablelabeling intensities were also observed in the medullaryreticular ®eld and in the inferior olive complex.

White matter areas throughout the brain showed littleor no labeling. Signal intensities in the lateral olfactorytract, the optic nerve, the anterior and posterior com-missure, the internal capsule, the sensory root of thetrigeminal nerve, and the pyramidal tract were not notice-ably higher than background (see Fig. 1 and below).Similarly, only background levels were detectable inthe choroid plexus and in ependymal cell layers. Like-wise, no signi®cant labeling was detected in a number ofnon-neural somatic tissues, including liver, lung, kidney,and spleen (data not shown).

A synopsis of the BC1 labeling intensities of majorbrain areas is given in Table 1. It had become apparentto us during the early stages of this work that there was asubstantial overlap between BC1-expressing neurons andsuch that express the mRNA for GAP-43. The distribu-tion of GAP-43 mRNA in selected brain areas haspreviously been analyzed by in situ hybridization.15,26

However, to ensure that expression patterns of GAP-43mRNA and BC1 RNA could be directly compared witheach other, it was necessary to analyze the distribution inbrain of both RNAs in parallel. Table 1 lists relativelevels of BC1 RNA labeling signals side-by-side withrelative levels of the GAP-43 mRNA labeling signalsin the same brain areas. In the following, the distributionof BC1 RNA in selected brain areas will be discussed, atthe cellular level, in comparison with the expressionpattern of GAP-43 mRNA.

Co-expression of BC1 RNA and growth-associatedprotein-43 messenger RNA in the rat CNS

A direct comparison of labeling patterns of BC1 RNAand GAP-43 mRNA is particularly informative in brainareas with lameinar organization. In the olfactory bulb,BC1 RNA and GAP-43 mRNA labeling patterns showeda characteristic overlap (Fig. 2). BC1 RNA was detect-able in the mitral cells and throughout the external plexi-form layer, with the density of silver grains decreasing inan outward gradient. The olfactory glomeruli, includingperiglomerular cells, were also labeled, although to alesser extent. The BC1 labeling signal in the inner ®elds

BC1/GAP-43 expression 467

Fig. 1. Distribution of dendritic BC1 RNA in rat brain. (A±G) BC1 expression patterns in a series of coronal sections from forebrain(A) through hindbrain (G). Labeling intensities are indicated by darkness of autoradiographic signal. Labeling intensities in whitematter areas (e.g. spinal trigeminal tract, sp5, in G) correspond to background labeling. (H) Coronal reference section probed for 7SLRNA. 7SL RNA, part of the signal recognition particle,38 is ubiquitously and non-preferentially expressed by all neurons and glial

cells. All abbreviations used in this and the following ®gures are, whenever possible, according to Paxinos.23

Table 1. Distribution of brain-speci®c cytoplasmic RNA1 and growth-associated protein-43 mRNA inthe adult rat nervous system

Labeling Intensities

Location BC1 RNA GAP-43 mRNA

TelencephalonOlfactory bulbInternal granular layer 1 1Mitral cell layer 111 111External plexiform layer* 11 ±/1 (tufted cells)Periglomerular cells 1 1Olfactory nerve layer ± ±Accessory olfactory bulb 111 111

Olfactory tubercle 1 1Anterior olfactory n 111 111Tenia tecta 11 11Cerebral cortexLayer 1* 11 ±/1Layers 2, 3 111 111Layer 4 11 11Layers 5, 6 111 111

HippocampusDentate gyrus

Granule cell layer ±/1 ±/1Polymorphic layer 111 111

CA1, CA2 11 11CA3 111 111Subiculum 11 11

Amygdaloid complexAnterior amygdaloid a 111 111Basolateral amygdaloid n 11 11Basomedial amygdaloid n 11 11Central amygdaloid n 111 11Medial amygdaloid n 111 11Medial amygdaloid n, posterodorsal 111 111Medial amygdaloid n, posteroventral 111 1 111 1Posteromedial cortical amygdaloid n 111 1 111Posterolateral cortical amygdaloid n 111 11Amygdalohippocampal a, posteromedial 111 111

Piriform cortex 111 111Dorsal endopiriform n 111 111Bed n of the stria terminalis 111 111Lateral septal n 111 111Caudate putamen 1 11Globus pallidus 1 11Ventral pallidum 1 11Nucleus basalis 11 11Substantia innominata 11 11

DiencephalonThalamusAnterodorsal thalamic n 111 111Anteroventral thalamic n 11 11Anteromedial thalamic n 111 111Centrolateral thalamic n 11 11Paraventricular thalamic n 111 111Laterodorsal thalamic n 111 11Ventrolateral thalamic n 11 11Ventromedial thalamic n 11 11Lateral posterior thalamic n 111 111Reuniens thalamic n 11 11Reticular thalamic n 11/111 1 /1 1Posterior thalamic nuclear group 11 1Ventral posteromedial thalamic n 11 1Ventral posterolateral thalamic n 11 1Lateral habenular n 111 11Dorsal lateral geniculate n 11 ndVentral lateral geniculate n 111 1 ndMedial geniculate n 111 nd

HypothalamusSupraoptic n² 111 1 1Medial preoptic a 111 111Lateral preoptic a 111 111Lateral hypothalamic a 111 111Posterior hypothalamic a 111 111Ventromedial hypothalamic n 111 111

Table 1 (continued)

Labeling Intensities

Dorsomedial hypothalamic n 111 111Zona incerta 11 11Arcuate n 11 111Median eminence 11 ±Premammillary n, ventral 111 111Lateral mammillary n 111 111

MesencephalonSubstantia nigra, compact 111 111Substantia nigra, reticular 11 1Superior colliculusZonal layer* 11 ±/1Super®cial gray layer 111 11Optical layer 11 11

Intermediate gray layer 11 11Olivary pretectal n 111 ndAnterior pretectal n 11 ndInferior colliculusDorsal cortex 111 11Central n 11 1

Dorsal raphe n 111 111Central gray 111 11Ventral tegmental a 111 111Ventrolateral tegmental a 11 11Deep mesencephalic n 111 11

Rhombencephalon (metencephalon/myelencephalon)Lateral and medial parabrachial n 111 111Ventral cochlear n, anterior 11/111 ndLocus coeruleus 111 111 1Pontine n 111 1Pontine raphe n 111 111Raphe pallidus n 1 1Medial vestibular n 11 1Gigantocellular reticular n 1 1Spinal trigeminal n 111 11Caudal spinal trigeminal n, gelatinous layer 111 1 111N solitary tract 111 1 111 1Inferior olive 111 111CerebellumMolecular layer* 1 ±Purkinje cell layer 1 /1 1 -/1Granular layer 1 /1 1 111

Cerebellar n² 111 1Spinal cord

Layer 1 11 11Layer 2 111 11Layers 3±6 11 11Layer 7 11 11Layers 8±10 1 1Dorsal corticospinal tract ± ±Ventral and lateral funiculus ± ±

Circumventricular organsSubfornical organ 111 111Area postrema 11 11

White matter areasOptic nerve ± ±Corpus callosum ± ±Anterior commissure ± ±Posterior commissure ± ±Fornix ±/1 ±Sensory root trigeminal nerve ± ±Pyramidal tract ± ±

Non-neuronal cellsChoroid plexus ± ±Ependymal cells ± ±Pineal gland ± ±

To establish correspondence of BC1 and GAP-43 labeling, signal intensities were determined overareas containing neuronal cell bodies. In neuropil areas (denoted by *), in contrast, GAP-43 labelingwas typically lower than BC1 labeling. We identi®ed two cell-body rich areas in which GAP-43labeling was substantially lower than BC1 labeling (denoted by ²); these areas may representexceptions to the rule of BC1/GAP-43 co-expression. Labeling intensities: ±, no signi®cant labeling;±/1, no signi®cant labeling in part of the area, with light labeling in other parts or subareas, or inspeci®c cells; 1, light labeling; 1 1 , moderate labeling; 111 strong labeling; 1111, verystrong labeling. n, nucleus/nuclei; a, area; nd, not determined.

of the glomeruli (the glomerular cores) is indicative of anextrasomatic location of the RNA. This was furthercon®rmed by in situ hybridization with 3H-labeled BC1probes (Fig. 2B, D). In comparison with 35S-probes, 3H-probes afford higher spatial resolution and little scatter,and the distribution of autoradiographic silver grainstherefore closely matches the actual location of targetBC1 RNA, in this case in glomerular cores (Fig. 2B).Similar to BC1 RNA, GAP-43 mRNA labeling wasstrongest in the mitral cells but silver grains were alsodetected over some individual cells, most likely tuftedcells, in the external plexiform layer. Neuropil in thislayer was largely devoid of speci®c GAP-43 labeling.Glomerular cores were not labeled, although a weaksignal was seen over periglomerular cells. For eitherRNA, only background levels were detectable over theolfactory nerve layer. These data indicate that BC1 RNAand GAP-43 mRNA were co-expressed in the same celltypes in the olfactory bulb (mitral/tufted cells and peri-glomerular cells). GAP-43 mRNA was mostly restrictedto neuronal cell bodies although a low-level presence inproximal-most dendritic segments can certainly not beruled out. In contrast to GAP-43 mRNA, BC1 RNAwas observed extending across entire dendritic ®elds ofthese cells, as is indicated by the BC1 labeling gradient inthe external plexiform layer and by the BC1 signal inglomerular cores. An analogous labeling pattern wasobserved in the accessory olfactory bulb (Fig. 2I±L).

Expression patterns of BC1 RNA and GAP-43 mRNAin the olfactory bulb were mirrored in hippocampus.Both RNAs were expressed at signi®cant levels in theAmmon's horn region (Table 1), with BC1 RNA presentthroughout dendritic ®elds and GAP-43 mRNA mostlyrestricted to cell bodies.35 In contrast to the Ammon'shorn region, BC1 and GAP-43 labeling signals were atvery low (but none the less above background) levels inthe layer of dentate granule cells (Table 1, Fig. 1). Verylow levels of GAP-43 mRNA in these cells have alsobeen described in earlier work.15 Low expression ofboth RNAs in the same neuronal cell type is in furthersupport of the notion of coordinate expression. A lowlabeling signal was detected for BC1 RNA (but not forGAP-43 mRNA) in the inner molecular layer of thedentate gyrus. This result indicates that BC1 RNA is infact targeted to proximal dendritic segments of dentategranule cells, albeit at low levels. Probes against bothRNAs strongly labeled scattered neurons in the poly-morphic layer (hilus) where GAP-43 mRNA labelingwas seen in distinct patches overlying individual neuronalcell bodies while BC1 labeling was more diffusely distrib-uted over somatic and neuropil areas (data not shown).

Analogous observations were made in other brainareas. Throughout the neocortex, laminar labelingpatterns were observed for both BC1 RNA and GAP-43 mRNA. For both RNAs, labeling intensities werehighest in layers II, III, V, and VI (Fig. 3). However, incontrast to GAP-43 mRNA, signi®cant BC1 labeling wasalso detectable in layer I, a neuropil layer that containsfew neuronal cell bodies. GAP-43 labeling was clusteredin neocortex, with silver grains overlying neuronalsomata. Similar patterns were observed in the piriform

cortex, the amygdaloid complex, and other allocorticalregions (Fig. 3, Table 1).

In the cerebellum, labeling for BC1 RNA was moder-ate over Purkinje cells (Fig. 4). An inward-to-outwarddecreasing labeling gradient was observed over the mole-cular layer, which contains dendritic arborizations ofPurkinje cells. GAP-43 mRNA labeling was low inPurkinje cells (data not shown). Relative expressionlevels of both RNAs were low to moderate in the cere-bellar granular layer, taking into account the very densepackaging of granule cells.

Both RNAs were found expressed at signi®cant levelsin the superior colliculus (Table 1). Unlike GAP-43mRNA, BC1 RNA was also detected at high levels inthe zonal layer (Fig. 4E, F), a neuropil layer with fewneuronal cell bodies. High levels of both RNAs were alsodetected in the inferior colliculus, in particular in thedorsal cortex. In contrast, neither RNA was detectablein the pineal gland (epiphysis). In the anterior ventralcochlear nucleus, the BC1 signal appeared highlyclustered over and surrounding neuronal somata (Fig.4A, C). In this area, ascending ®bers of the auditorynerve form large synaptic endings (the end bulbs ofHeld) which terminate on the somata of secondaryneurons.39 Typically, such a synaptic ending completelyengulfs the soma of the secondary neuron (e.g. a bushycell) onto which it terminates. Therefore, the clusteredsomatic distribution of BC1 RNA in the ventral cochlearnucleus is likely to be a re¯ection of the axo-somaticrather than axo-dendritic synaptology of these neurons.

In the brainstem, BC1 RNA and GAP-43 mRNA weredetected at high levels in the locus coeruleus, the nucleusof the solitary tract, nuclei of the inferior olive, the dorsalmotor nucleus of vagus, the spinal trigeminal nucleus,and others (Fig. 5, Table 1, and data not shown). As inother brain areas, silver grains indicating the presence ofGAP-43 mRNA were seen overlying nerve cell bodieswhereas the BC1 signal extended into neuropil areas. Inthe spinal cord, labeling for BC1 RNA and GAP-43mRNA was highest in the dorsal horn, in particular inlamina 2 (not shown). GAP-43 mRNA labeling wasagain predominantly somatic while the BC1 RNA wasseen extending into neuropil areas. White matter was notlabeled above background levels with either BC1 RNAor GAP-43 mRNA probes.

DISCUSSION

BC1 RNA has been implicated in the regulation oflocal protein synthesis in postsynaptic microdomains inneurons,3 a mechanism now recognized to play an impor-tant role in the implementation of long-term synapticplasticity.12,29,32 However, such mechanisms are not likelyto be implemented uniformly in all neuronal subtypes asvarious types of neurons can be expected to differ fromeach other in their capacity to modulate postsynapticform and function. It was therefore necessary to establishthe expression pattern of BC1 RNA in the nervoussystem, and to cross-differentiate this pattern againstother patterns of functional neuronal heterogeneity.

BC1/GAP-43 expression 471

Y. Lin et al.472

Fig. 2 (A±H).

We report here that BC1 RNA is expressed by adistinct subset of neurons, and that this subset does notcoincide with any particular neuronal cell type, transmitter-receptor system, or other cellular and regional patterns inthe nervous system. However, we found that BC1-expressing neurons typically also express the mRNAencoding the axonal phosphoprotein GAP-43. In matureneurons, expression of GAP-43 has been shown to be adeterminant of neuronal plasticity.2 In conclusion, ourresults thus identify co-expression of BC1 RNA andGAP-43 mRNA as a molecular hallmark of a subset ofmature and highly plastic CNS neurons.

Although co-expressed in various neuronal subtypes,the subcellular distribution of BC1 RNA and GAP-43mRNA within such neurons is strikingly different:somatodendritic for BC1 RNA, predominantly somaticfor GAP-43 mRNA. This differential subcellular distri-bution, although apparent throughout the nervoussystem, is best observed in strati®ed areas such as olfac-

tory bulb, retina,34,35 neocortex or hippocampus. In theolfactory bulb, BC1 RNA is detected not only in mitralcell bodies, but also in a decreasing gradient over theentire external plexiform layer, indicating that the RNAis located over the entire mitral/tufted dendritic extent inthis layer. Dendrodendritic synapses (between mitral/tufted lateral dendrites and granule cell dendrites) areprominent in this layer, as they also are in the internalplexiform layer of the retina where BC1 RNA isexpressed at equally high levels.34,35 These data thereforesuggest that high levels of BC1 RNA are targeted to thisparticular type of synapse.

Signi®cant levels of BC1 RNA are furthermore presentinside olfactory glomeruli. The fact that glomerular coresshow a positive signal also with 3H-labeled BC1 probescon®rms that the signal within these ®elds is genuineand not attributable to scatter from periglomerularcells. It is within glomeruli that axons of olfactory recep-tor neurons terminate on mitral apical dendrites and on

BC1/GAP-43 expression 473

Fig. 2. Expression of BC1 RNA and GAP-43 mRNA in the olfactory bulb. (A±H) Main olfactory bulb. (A) BC1 RNA, 35S-label; (B)BC1 RNA, 3H-label. (C) GAP-43 mRNA, 35S-label. (D) Bright-®eld photomicrograph corresponding to B. BC1 labeling extendsfrom the layer containing mitral cell bodies (arrows in B and D) in a decreasing gradient into the external plexiform layer (EPL).BC1 detection with a 3H-probe shows that cores of glomeruli are also labeled (B, D; GL, glomerular layer). Little labeling isobserved in the olfactory nerve (ON). Labeling for GAP-43 mRNA is restricted to the mitral cell body layer (C, arrow), tufted cells inthe EPL (C, arrowhead), and periglomerular cells, but in contrast to BC1 RNA not in neuropil areas. (E±H) Labeling patterns of BC1RNA (E, F) and GAP-43 mRNA (G, H) at higher magni®cations. Some clusters of silver grains that indicate labeled somata in thelayer containing mitral cell bodies are identi®ed by arrows. The distribution of the labeling signals indicates that GAP-43 RNA isexpressed in the same cells as BC1 RNA but remains mostly restricted to the somata of those cells. (I±L), Accessory olfactory bulb.35S-labeling for BC1 RNA (I) is diffuse in the layer of output neurons (O) whereas labeling for GAP-43 mRNA (K) is clustered overcell bodies. BC1 detection with a 3H-probe shows labeling in cores of glomeruli (J). (L) Bright-®eld photomicrograph corresponding

to J. Scale bars� 100 mm (A±D, I±L), 50 mm (E, G), 25 mm (F, H).

Fig. 2 (I±L).

Y. Lin et al.474

Fig. 3. Distribution of BC1 RNA (A, C, E, G) and GAP-43 mRNA (B, D, F, H) in cerebral cortex. (A±D) Primary somatosensorycortex. Cortical layers are indicated by roman numerals. BC1 signal is signi®cant in layer I; in contrast, little GAP-43 signal isdetected in this layer. Pir (E±H): BC1 labeling is diffuse, consistent with a somatodendritic location; GAP-43 labeling is concen-trated over neuronal cell bodies and does not signi®cantly extend into the outermost cortical layer (external plexiform layer). Bright-

®eld photomicrographs are shown underneath corresponding dark-®eld photomicrographs. rf, rhinal ®ssure. Scale bar� 100 mm.

periglomerular cell dendrites. BC1 labeling in glomeru-lar cores could thus be due to the presence of BC1 RNAin mitral or periglomerular cell dendrites, in nerve term-inals of olfactory receptor neuron axons, or both. Olfac-tory receptor neuron axonal terminals have been shownto contain odorant receptor mRNAs,25,37 but it is unclearwhether such mRNAs are locally translated, and if so,whether such locally translated axonal odorant receptorsmight play a role in the path®nding and/or target recog-

nition of axonal growth cones in proliferating olfactoryreceptor neurons. Should BC1 RNA in fact be axonal inthis system, it is conceivable that it is involved in thelocal regulation of axonal translation. It should bepointed out, however, that the absence of speci®cBC1 labeling over the olfactory nerve indicates that littleif any BC1 RNA is present in axonal shafts ofthese neurons. While not directly ruling out thepresence of BC1 RNA in the terminals of such axons,

BC1/GAP-43 expression 475

Fig. 4. BC1 expression in cerebellum, superior and inferior colliculus, and anterior ventral cochlear nucleus. In the Cb, BC1 RNA isexpressed at moderate levels in Purkinje cell somata (arrows in A, dark ®eld; C, bright ®eld) and in a decreasing gradient across themolecular layer (CbMol). BC1 RNA is also detectable in the cerebellar granule cell layer (CbGra) although this is partially obscured bythe dense packaging of Cresyl Violet-stained granule cell somata. Longer exposure times were used in A to compensate for theoverall moderate BC1 expression in Cb; compare with standard exposure in B (sagittal section; Cb). (A, C) BC1 labeling signal isstrong and strikingly clustered in the VCA. (B, D) BC1 expression is equally strong in the DCIC, but not detectable in the pinealgland (Pi) or chp. In the superior colliculus (E, dark ®eld; F, bright ®eld), in situ hybridization with a 3H-labeled probe shows thatBC1 RNA is localized to the Zo, a neuropil layer superjacent to the super®cial gray layer (SuG). Scale bar� 100 mm (A±D), 50 mm

(E, F).

Y. Lin et al.476

Fig. 5. BC1 RNA and GAP-43 mRNA in brainstem. In all areas analyzed, expression of BC1 RNA (left column, A, C, E, G, I, K) and GAP-43 mRNA(right column, B, D, F, H, J, L) overlap such that the latter appears clustered over cell bodies while the former is observed over cell bodies andadjacent neuropil areas. Such differential distribution, indicating co-expression but different subcellular localization of both RNAs, is apparent in theLC, the nucleus of the solitary tract (Sol) including the commissural part (SolC), and the inferior olive (IOA, IOB, IOC: subnuclei A, B, C). Bright-®eld photomicrographs are shown underneath corresponding dark-®eld photomicrographs. AP, area postrema; CC, central canal; py, pyramidal tract.

Scale bars� 100 mm.

this observation is more in line with a predominantlydendritic nature of the BC1 signal in glomerular cores.Clearly, higher resolution techniques will be needed in thefuture to address the possibility of BC1 targeting to theseand other axons.

The two characteristics of BC1 expression (co-distri-bution with GAP-43 mRNA in neuronal somata, andextension along dendritic extents), as exempli®ed in theolfactory bulb, were similarly observed in other regionsthroughout the brain, including strati®ed regions such asneocortex, superior colliculus and cerebellum, as well asin areas with less obvious neuropil layering (see Table 1for a comprehensive listing). While GAP-43 mRNAlabeling was typically seen clustered over neuronalsomata, our data do not rule out the possibility thatlower amounts of the RNA are actually present inproximal-most dendritic segments.

It should be noted that the expression of dendritic BC1RNA in the nervous system, while characteristically co-inciding with that of GAP-43 mRNA, is not congruent withexpression patterns of prototypical dendritic mRNAs.Dendritic MAP2 mRNA, for example, was identi®ed inproximal dendrites22 whereas BC1 RNA is detectablealong the entire dendritic extent. Other dendritic mRNAs,for example those encoding CaMKIIa5 and the IP3 recep-tor type 1,9 are expressed in a more region-selectivemanner (forebrain in the former, cerebellum in the lattercase) as compared with BC1 RNA. Evidently, morerestricted expression patterns are direct re¯ections of thefunctions of the respective proteins encoded by these

dendritic mRNAs. The more widespread distribution ofBC1 RNA, in contrast, supports the notion of a moregeneral role of this RNA in neuronal plasticity.

In summary, the results presented in this paper indicatethat BC1 RNA is located in postsynaptic microdomainsin a distinct subset of neurons in the rat nervous system.Depending on the BC1-expressing cell type, such post-synaptic location can manifest itself as predominantlysomatic (bushy cells), somatodendritic (pyramidalcells), or dendritic (granule cells) in nature. Consistentwith our results is earlier biochemical work7 whichshowed that BC1 RNA is the most prominent RNAspecies in spine preparations from CA3 pyramidal cellapical dendrites. The combined data therefore support arole of BC1 RNA in postsynaptic RNA metabolism andmodulation.

The second major ®nding of this report is the wide-spread co-expression of BC1 RNA and GAP-43 mRNAin various neuronal subtypes. In mature neurons, GAP-43has been implicated in presynaptic,2 BC1 RNA in post-synaptic plasticity.3 GAP-43/BC1 co-expression thereforeimplies orchestration of gene expression within the sameneuron to modulate throughput of synaptic trans-mission17 at both pre- and postsynaptic sites, thus identi-fying neurons that display simultaneous competence forboth types of plasticity. Moreover, our data are in agree-ment with the notion that both pre- and postsynapticmechanisms contribute to synaptic change, a view thathas earlier been discussed with reference to GAP-43function.8 Further support comes from the observation

BC1/GAP-43 expression 477

Fig. 5 (I±L).

that presynaptic terminals of the mossy ®berÐCA3synapse in the stratum lucidum of hippocampus lackGAP-43 (reviewed by Colley and Routtenberg8). Thesesynapses also display a different form of long-termpotentiation, one that is not dependent on N-methyl-d-aspartate receptors.8 At the same time, levels of BC1RNA have been observed to be signi®cantly lower instratum lucidum segments of CA3 apical pyramidalcell dendrites than in stratum radiatum segments.35

Thus, GAP-43 and BC1 RNA are modulated coordi-

nately along the same dendrite, the former in presynapticlocation, the latter in postsynaptic. An important issue forfuture work (preferably including ultrastructural analy-sis) will therefore be the question whether such coordi-nate modulation of BC1 RNA and GAP-43 at the synapseis characteristic of other types of synapses as well.

AcknowledgementsÐWe thank R. Neve for the clone used togenerate GAP-43 probes. This work was supported by NationalInstitutes of Health grant NS34158 (to H.T.).

REFERENCES

1. Bassell G. J., Oleynikov Y. and Singer R. H. (1999) The travels of mRNAs through all cells large and small. Fedn Proc. Am. Socs exp. Biol. 13,447±454.

2. Benowitz L. I. and Routtenberg A. (1997) GAP-43: an intrinsic determinant of neuronal development and plasticity. Trends Neurosci. 20,84±91.

3. Brosius J. and Tiedge H. (2001) Dendritic BC1 RNA: intracellular transport and activity-dependent expression. In Cell polarity and subcellularRNA localization (ed. Richter D.), Springer, Heidelberg, pp. 129±138.

4. Bruckenstein D. A., Lein P. J., Higgins D. and Fremeau R. T. (1990) Distinct spatial localization of speci®c mRNAs in cultured sympatheticneurons. Neuron 5, 809±819.

5. Burgin K. E., Waxham M. N., Rickling S., Westgate S. A., Mobley W. C. and Kelly P. T. (1990) In situ hybridization histochemistry of Ca21/calmodulin-dependent protein kinase in developing rat brain. J. Neurosci. 10, 1788±1798.

6. Cheng J.-G., Tiedge H. and Brosius J. (1996) Identi®cation and characterization of BC1 RNP particles. DNA Cell Biol. 15, 549±559.7. Chicurel M. E., Terrian D. M. and Potter H. (1993) mRNA at the synapse: analysis of a preparation enriched in hippocampal dendritic spines.

J. Neurosci. 13, 4054±4063.8. Colley P. A. and Routtenberg A. (1993) Long-term potentiation as synaptic dialogue. Brain Res. Rev. 18, 115±122.9. Furuichi T., Simon-Chazottes D., Fujino I., Yamada N., Hasegawa M., Miyawaki A., Yoshikawa S., GueÂnet J.-L. and Mikoshiba K. (1993)

Widespread expression of inositol 1,4,5-trisphosphate receptor type 1 gene (Insp 3r 1) in the mouse central nervous system. Receptors andChannels 1, 11±24.

10. Gardiol A., Racca C. and Triller A. (1999) Dendritic and postsynaptic protein synthetic machinery. J. Neurosci. 19, 168±179.11. Kiebler M. A. and DesGroseillers L. (2000) Molecular insights into mRNA transport and local translation in the mammalian nervous system.

Neuron 25, 19±28.12. Kindler S., Mohr E. and Richter D. (1997) Quo vadis: extrasomatic targeting of neuronal mRNAs in mammals. Molec. Cell. Endocrinol.

128, 7±10.13. Kleiman R., Banker G. and Steward O. (1993) Subcellular distribution of rRNA and poly (A) RNA in hippocampal neurons in culture. Molec.

Brain Res. 20, 305±312.14. Kobayashi S., Goto S. and Anzai K. (1991) Brain-speci®c small RNA transcript of the identi®er sequences is present as a 10 S ribonucleoprotein

particle. J. biol. Chem. 266, 4726±4730.15. Kruger L., Bendotti C., Rivolta R. and Samanin R. (1993) Distribution of GAP-43 mRNA in the adult rat brain. J. comp. Neurol. 333, 417±434.16. Landry C. F., Watson J. B., Kashima T. and Campagnoni A. T. (1994) Cellular in¯uences on RNA sorting in neurons and glia: an in situ

hybridization histochemical study. Molec. Brain Res. 27, 1±11.17. Meberg P. J., Barnes C. A., McNaughton B. L. and Routtenberg A. (1993) Protein kinase C and F1/GAP-43 gene expression in hippocampus

inversely related to synaptic enhancement lasting 3 days. Proc. natn. Acad. Sci. U.S.A. 90, 12,050±12,054.18. Mohr E. (1999) Subcellular RNA compartmentalization. Prog. Neurobiol. 57, 507±525.19. Muslimov I. A., Banker G., Brosius J. and Tiedge H. (1998) Activity-dependent regulation of dendritic BC1 RNA in hippocampal neurons in

culture. J. Cell Biol. 141, 1601±1611.20. Muslimov I. A., Santi E., Homel P., Perini S., Higgins D. and Tiedge H. (1997) RNA transport in dendrites: a cis-acting targeting element is

contained within neuronal BC1 RNA. J. Neurosci. 17, 4722±4733.21. Neve R. L., Finch E. A., Bird E. D. and Benowitz L. I. (1988) Growth-associated protein GAP-43 is expressed selectively in associative regions

of the adult human brain. Proc. natn. Acad. Sci. U.S.A. 85, 3638±3642.22. Paradies M. A. and Steward O. (1997) Multiple subcellular mRNA distribution patterns in neurons: a nonisotopic in situ hybridization analysis.

J. Neurobiol. 33, 473±493.23. Paxinos G. and Watson C. (1998) The Rat Brain in Stereotaxic Coordinates. Academic, San Diego.24. Rao A. and Steward O. (1993) Evaluation of RNAs present in synaptodendrosomes: dendritic, glial, and neuronal cell body contribution.

J. Neurochem. 61, 835±844.25. Ressler K. J., Sullivan S. L. and Buck L. B. (1994) Information coding in the olfactory system: evidence for a stereotyped and highly organized

epitope map in the olfactory bulb. Cell 79, 1245±1255.26. Rosenthal A., Chan S. Y., Henzel W., Haskell C., Kuang W.-J., Chen E., Wilcox J. N., Ullrich A., Goeddel D. V. and Routtenberg A. (1987)

Primary structure and mRNA localization of protein F1, a growth-related protein kinase C substrate associated with synaptic plasticity. Eur.molec. Biol. Org. J. 6, 3641±3646.

27. Schuman E. M. (1999) mRNA traf®cking and local protein synthesis at the synapse. Neuron 23, 645±648.28. Skene J. H. P. (1989) Axonal growth-associated proteins. A. Rev. Neurosci. 12, 127±156.29. Steward O. (1997) mRNA localization in neurons: a multipurpose mechanism? Neuron 18, 9±12.30. Steward O. and Reeves T. M. (1988) Protein-synthetic machinery beneath postsynaptic sites on CNS neurons: association between polyribo-

somes and other organelles at the synaptic site. J. Neurosci. 8, 176±184.31. Tiedge H. (1991) The use of UV light as a cross-linking agent for cells and tissue sections in in situ hybridization. DNA Cell Biol. 10, 143±147.32. Tiedge H., Bloom F. E. and Richter D. (1999) RNA, Wither Goest Thou? Science 283, 186±187.33. Tiedge H. and Brosius J. (1996) Translational machinery in dendrites of hippocampal neurons in culture. J. Neurosci. 16, 7171±7181.34. Tiedge H., DraÈger U. C. and Brosius J. (1992) Murine BC1 RNA in dendritic ®elds of the retinal inner plexiform layer. Neurosci. Lett. 141,

136±138.

Y. Lin et al.478

35. Tiedge H., Fremeau R. T. Jr, Weinstock P. H., Arancio O. and Brosius J. (1991) Dendritic location of neural BC1 RNA. Proc. natn. Acad. Sci.U.S.A. 88, 2093±2097.

36. Torre E. R. and Steward O. (1996) Protein synthesis within dendrites: glycosylation of newly synthesized proteins in dendrites of hippocampalneurons in culture. J. Neurosci. 16, 5967±5978.

37. Vassar R., Chao S. K., Sitcheran R., Nunez J. M., Vosshall L. B. and Axel R. (1994) Topographic organization of sensory projections to theolfactory bulb, Cell 79, 981-991.

38. Walter P. and Blobel G. (1982) Signal recognition particle contains a 7S RNA essential for protein translocation across the endoplasmicreticulum. Nature 299, 691±698.

39. Webster W. R. (1995) Auditory system. In The Rat Nervous System (ed. Paxinos G.), pp. 797±831. Academic, San Diego.

(Accepted 20 December 2000)

BC1/GAP-43 expression 479