Neuron, Vol. 14, 433-445, February, 1995, Copyright © 1995 by Cell Press Arc, a Growth Factor and Activity-Regulated , Encodes a Novel C oskeleton-Associated Protein That Is Enriched in Neuronal Dendrites

Gregory L. Lyford,* 1 Kanato Yamagata,*l sistent with classical studies of and that Walter E. Kaufmann,t I Carol A. Barnes,~ demonstrate a requirement for protein synthesis in long- Laura K. Sanders,§ Neal G. Copeland,ll term, but not short-term, memory (Flexner et al., 1963; Debra J. Gilbert,II Nancy A. Jenkins, II Agranoff, 1981; Davis and Squire, 1984). Anthony A. Lanahan,* and Paul F. Worley*t Insight into genomic mechanisms that might underlie *Department of Neuroscience long-term plasticity initially came from studies of growth 1"Department of Neurology factor signaling in nonneuronal tissues. Growth factor §Howard Hughes Medical Institute stimulation induces the rapid and transient expression of and Department of Molecular Biology and a set of , termed immediate-early genes (lEG), that Johns Hopkins University School of Medicine encode transcription factors and cytokines, as well as Baltimore, Maryland 21205 other molecules, that are believed to regulate long-term tDepartment of Psychology and Neurology cellular responses (reviewed by Lau and Nathans, 1991). and Division of Neuronal Systems, Memory, and Aging Similar rapid genomic responses are induced in University of Arizona by neurotransmitter stimulation (Greenberg et al., 1986; Tuscon, Arizona 84724 Sheng and Greenberg, 1990). Kandel and coworkers have IIMammalian Genetics Laboratory used a simplified cultured system of invertebrate neurons ABL-Basic Research Program to demonstrate an essential role for both mRNA and pro- NCI-Frederick Cancer Research and Development Center tein synthesis in long-term synaptic plasticity (Montarolo Frederick, Maryland 21702 et al., 1986) and a direct role for the lEG CCAAT enhancer-binding protein (Alberini et al., 1994). Summary To identify components of the genetic program underly- ing long-term responses in the vertebrate , we and Neuronal activity is an essential stimulus for induction others have used differential cloning techniques to identify of plasticity and normal development of the CNS. We mRNAs that are rapidly induced by excitatory activity have used differential cloning techniques to identify a (Nedivi et al., 1993; Qian et al., 1993; Yamagata et al., novel immediate-early gene (lEG) cDNA that is rapidly 1993, 1994a, 1994b). In addition to transcription factors induced in neurons by activity in models of adult and (Yamagata et al., 1994a), this approach has identified a developmental plasticity. Both the mRNA and the en- number of lEGs that encode enzymes that may directly coded protein are enriched in neuronal dendrites. modify cellular function, including tissue plasminogen acti- Analysis of the deduced amino acid sequence indi- vator (Qian et al., 1993), an inducible form of cyclooxygen- cates a region of homology with ~x-spectrin, and the ase (Yamagata et al., 1993), and a novel homolog of H-Ras full-length protein, prepared by in vitro transcription/ (Yamagata et at., 1994b). These proteins presumably in- translation, coprecipitates with F-actin. Confocal mi- teract with existing cellular proteins to modify properties croscopy of the native protein in hippocampal neurons of the and effect long-term changes in cellular func- demonstrates that the lEG-encoded protein is en- tion. Here, we describe a novel lEG, which encodes a riched in the subplasmalemmal codex of the cell body protein that exhibits homology to the structural protein and dendrites and thus colocalizes with the actin cy- ~-spectrin and is localized to neuronal dendrites. We have toskeletal matrix. Accordingly, we have termed the termed the gene Arc (activity regulated cytoskeleton- gene and encoded protein Arc (activity-regulated cy- associated protein). Because its mRNA and protein ex- toskeleton-associated protein). Our observations sug- pression are tightly coupled to the activity state of the gest that Arc may play a role in activity-dependent neuron, we hypothesize that Arc may play a role in activity- plasticity of dendrites. dependent changes in dendrite function.

Introduction Results

Recent studies support the concept that neuronal plastic- Arc cDNA Sequence ity involves two temporally distinct components (Goelet et A novel cDNA corresponding to the 3' noncoding region al., 1986; Nguyen et al., 1994). Transmitter stimulation of a - 3.2 kb m RNA was identified by differential screening results in rapid membrane ionic conductance changes and of a subtracted cDNA library prepared from seizure- associated protein phosphorylation events that underlie stimulated hippocampus (Yamagata et al., 1993). A near short-term cellular changes lasting minutes to hours. By full-length cDNA was identified by iterative screening of contrast, long-term plasticity is distinguished from short- a brain cDNA library prepared in our laboratory. The size term plasticity in that it requires the synthesis of new of the cDNA (3032 bp; Figure 1) corresponds closely to mRNA and proteins. This mechanistic framework is con- the estimated size of the mRNA as determined by Northern analysis. The longest open reading frame (ORF) predicts 1These authors made nearly equivalent contributionsto this paper. a 396 amino acid protein. Translation stop sites are pres- Neuron 434

~T~TGG~GAGTA~T~CT~ccT~A~cc~AGT~7~c~c~T~TT~AG~TTc~ccc~GA~G~A~A~Tcc~G~T 90 ~A~G~A~TT~A~T~T~c~TT~G~c~AG~G~T~TG~T~TTcTG~TTc~c~cAGAGGAGT 180 translated and coding regions. TNT of the full length cDNA 0 10 HELDH~TTGGL~AYpApR TCTTAGCCTGTCcCG~CC~T~CCCC~CGAGCAGAT~GAGCTGGACCATA~ACGACCGGcGGCCTCCAC~CTACCCTGcCCC~ 270 resulted in synthesis of a protein that migrated on SDS- 2O 30 4O ~Gp~AKPNVILOIGKCRAE~V~T*H~ polyacrylamide gel electrophoresis (SDS-PAGE) with an SO ~0 ~0 apparent molecular weight of 55 kDa (see Figure 9). TNT of

VPTGDS'QRWKKS*IKACLCRC~ETIANLERW cDNAs restricted at sites within the predicted ORF yielded products of reduced size, consistent with the position of ~TcAAGcGTGA~ATGcAcG~GT~As~A~T~TTcTA~sT~TGGAGAsGT~c~Acc~cTGsAGT~AT~GGc~GTAcccA 63O restriction sites (data not shown). Additionally, TNT of a GTGGG~AG~GA~G~CC~CACACTGTCT~TGTAGGTG~G~T~CAGAG~TA~T~CAGG~T~ATGGCTAC~TA~A~T 720 cDNA fragment lacking the putative translation start site VSPYAITpppAAGELP~QESVGAQQYQS+WV GTTA~ccTAT~A~A~CC~G~A~CT~c~GcAGGA~AG~TG~cTGA~A~GTcAGTTG~G~T~Asc~A~AGTcTTGsGTG 810 but containing other 3' methionines yielded no product. PGEDSQPSPGLDTQXFEDPR~FLSHL~EYL ccAGs~AGGA~G~c~c~=cccAGGTcT~ATAcccA~AT~T~AssAcc~Ae~GAsTTcc~sAGccAcCT~GAsTAc~T~ 900 Further confirmation that the identified ORF is correct was obtained by generating antisera against a bacterial fusion protein that recognizes an inducible protein in hippocam- ~TcGGTG~s~cTGsGTGGAGTTc~G~AGTTzcTscAGTAcA~TGA~GTAc~T~T~c~cG~cAT~AGC~GGA~Ts~8~

DLpQKQGEpLDQFLW~KRDLYQTLYVDAEE pus that is very similar in size to the TNT product (see GAc~TGccAcAG~A~G~GA~ccAcTT~A~cAGT~cTcT~GT~c~GAcc~YaccAGAcAc~GTATGT~Ac~CT~A~As I170

EEIIQYVVGTLQPKFKRFLRHPLFKTLEQL below). GAGGAGA~A~CAGTATGT~TG~cAC~TGcAG~C~GTTC~G~TTT~TGCGCCACCCACTT~Cc~GACCCTCGAGCA~TC 1260 We compared the predicted Arc amino acid sequence IQRGMEVQDGLEQAAEPSVTPLPTEDETEA ATCCAGA~GGCATGG~GTTCAGGAC~CCTGGAGCAG~A~TGAGCCTTCTGTCACCCCTCTGCCCACAGAGGATGAGACTGA~CA]35~ ~0 s~o with the GenBank database to identify homologous pro-

CTCACGCCTGCTCTTACCAGCCAGTCAGTAGCCAGTGACAG~ACCCAGCCTG~TAGA~GCCAGCCCAGGGTCCCCAGCCTGCCTGCC 1440 ACACCCAGTCTGTGGCTT¢TGTC~CTACGACTTGATTGAG~T~GGCT~CACCC~ATGCCCTCTCCAGCCAGACACCTTCTCA 1530 teins. The closest homology identified is with ~-spectrin

CCTCACCTGT~TG~CCTAGTCCTG~CCTGTCTCCAGT~GCCTCACCCTCTACACTCTCAGACCATCACAG~CACCTTT~CT~CTCA 1710 (Wasenius et al., 1989; Moon and McMahon, 1990) (Fig- T TC T~ATCAGTGTCCAGG~CC CTTTG~ TAGTC~G~TC~CTGTCTG~GGC~G TAG~ACC~C C~GGC~A T C 1800 ure 2). The region of homology spans 156 amino acids in CGCT~C~CTCCT~CCCCCAGCAGTGATTCATACCAGT~G~GCAGACTTCGGCTCCATGACTCAGCCATGCCA~GCGGA~GT 2070 CCCAGA~GCTGAGTCC~AGCCCCA~TGA~CAGCAGCT~AGTCT~AGAGCC~TG~TGACACCA~TCTC~TGCTCAGA 2160 the carboxyl terminal half of Arc (amino acids 155-316), CACTTTC~AGAG~TAGGATTAC~GA~AGGTGCCTTGGCA~G~CAGCACCCAGCTCA~CTCAGA~TG~GGTG~GA 2340 C~GCCA~GTG~CCCC~GTCTG~CACG~T~CC¢CTCCGCTGGCCAC~ACCA~T~CTGCCAC~GCCACT~AGCTTGAGCA243~ where it is 20% identical to the twenty-first and twenty-

CACTTCCCTGACCTGCC~G~AT~CCCAGCT~CC~ATCCACACCCCAGCACC~CCACCTCGT~TTGGT~CTGCTCGTGTCTG 2700 second repeats of ~-spectrin. Though this level of identity is generally too low to be predictive of function, we noted that the degree of identity between repeats in c~-spectrin is Figure 1. Nucleotide and Amino Acid Sequences of Arc typically only 20% (Wasenius et al., 1989). Each a-spectrin Numbers to the right refer to last nucleotide in each line. Predicted repeat is believed to form a tri-a-helix bundle that is stabi- amino acid sequence is displayed above corresponding nucleotide lized by hydrophobic, leucine zipper-like interactions with sequence, and numbersfor amino acids are indicatedabove. Asterisk neighboring coils (Yan et al., 1993). Amino acids that are indicates predicted phosphorylation sites for protein kinase C; plus sign indicates predicted phosphorylation sites for calcium/calmodulin conserved between spectrin repeats correspond to sites dependent kinase I1. ~o%T~0% polyadenylation signal and ATIrA of protein-protein interaction between coils. Because mRNA instability signals are underlined. many of the amino acid differences between a-spectrin and Arc represent conservative changes, the overall level of homology in this region is 77%. Accordingly, the de- duced amino acid sequence of Arc suggests that it pos- ent in all three reading frames 5'to the putative translation sesses a spectrin-like domain that may be important in start site, and the nucleotide sequence surrounding the forming intermolecular interactions. A separate search of predicted initiating methionine conforms to the Kozak con- GenBank using only the NH2 terminal half of Arc was unin- sensus for efficient translation, including a 5' GC-rich re- formative. gion (Kozak, 1987, 1989). Methionines are also present in amino acid positions 6 and 36 of this ORF but are not Arc mRNA Is Enriched in Brain and Is Rapidly optimal for translation initiation. The 3' noncoding region Induced by Neuronal Activity and Growth Factors includes a single ATTTA sequence that has been imp!i- ' We examined the expression and regulation of Arc mRNA cated in destabilizing mRNAs (Shaw and Kamen, 1986) in various tissues and in cultured PC12 cells. The 3.2 kb and a polyadenylation signal (AATAAA) 18 nucleotides message was detected in hippocampus and was strongly from the 3' terminal poly(A) tail. induced within 30 min after an electrically induced seizure (maximal electroconvulsive seizure [MECS]; Figure 3A). Predicted Protein Sequence: Arc mRNA levels remain elevated for 8 hr after MECS Homology to a-Spectrin and return to basal levels by 24 hr. Arc mRNA was also The predicted protein has a calculated molecular weight detected in normal cerebral cortex and rapidly increased of 45,365 Da, a pl of 4.5, and is generally hydrophilic, with following MECS. Basal expression of Arc mRNA in hippo- no identified signal sequence, no hydrophobic stretches campus and cerebral cortex rapidly decreased following of sufficient length to be membrane spanning, and no pre- administration of the N-methyI-D-aspartate receptor an- dicted glycosylation sites. Several possible phosphoryla- tagonist MK-801, suggesting that the level of basal expres- tion sites were identified, including sites for protein kinase sion is regulated by natural excitatory synaptic activity, as C and calcium/calmodulin-dependent kinase 2 (legend to previously demonstrated for the transcription factor lEG Figure 1). zif268 (Worley et al., 1991). Arc mRNA expression is strik- To confirm the predicted ORF, we examined the size ingly enriched in brain relative to peripheral tissues. of the in vitro transcription/translation (TNT) product pre- The developmental profile of Arc mRNA expression was pared from the full-length cDNA, as well as from cDNAs determined in rat forebrain (Figure 3B). Arc mRNA was that were truncated by restriction in the predicted 3' non- first detected at postnatal day 8 and subsequently in- Arc Expressionand Regulationin Brain 435

* * ** . Figure 2. Comparisonof the Rat Arc Amino spectrin 2190 LQKIIKEREL ELQKEQR/~QE ENDKLRQEFA QHANAFHQWI QETRTYLLDG 2239 (repeat 21) I .... I I : i ..... I : II t :: ::I: ..... I , Acid Sequencewith Repeats 21 and 22 of Arc 228 LRQVGGSEEY WLSQIQNHMN GPAKKWWEFK Q--GSVKNWV EFKKEFLQYS 275 Chicken ~-Spectrin L Q G EE L Q QNH N AK EFK Q G VK W E KKE LQ S Amino acid residues228-380 of Arc aligned with residues2190-2337 of a-spectrin. Spec- spectrin 2240 SCMVEESGTL ESQLEATKRK HQEIRA---M RSQLKKIEDL GAAM~EAL-- 2284 trin residues 2190-2284 are from repeat 21, (repeat 21) ...... I ...... I .... I: IE: Arc 276 EGTL-SREAI QRELDLPQKQ GEPLDQRLQR KRDLYQTLYV DAEEEEIIQY 324 and residues2285-2337 are from repeat 22. GTL REA R L QKQ G L Q K DLYQT AEEEE A bar indicatesidentity between Arc and spec- * * * ** trin; two dots indicateconservative changes. spectrin 2285 ---ILDNKYT EHSTVGLAQQ WDQLDQLGMR MQHNLEQQIQ ARNTTGVTEE 2331 Aminoacids that are conservedbetween multi- (repeat 22) :I: I:: I;: :I[ I II :I::II[ .... I: II: ple repeatsof spectrin are indicatedwith an Arc 325 WGTLQPKFK RFLRHPLPKT LEQLIQRGME VQDGLEQAAE PSVTPLPTED 374 TLQ K R L KT LEQL Q GME Q DLEQA E S T TE asterisk (Wasenius et al., 1989), and amino acids that are present in at least one repeatof spectrin are indicatedbelow the alignmentin spectrin 2332 ALKEFS 2337 (repeat 22) : :::: italics. Arc 375 ETEALT 380 E EAL

creased to peak levels of expression at postnatal day 21. stimulus was similar to that produced by MECS in the Arc mRNA continues to be expressed at relatively high hippocampus. The magnitude of Arc mRNA induction by levels in the adult forebrain. This pattern of developmental the HF stimulus was also comparable to that following expression parallels that of several other lEGs, including MECS; however, unlike MECS, which induces Arc mRNA zif268 (Worley et al., 1990; Kaufmann et al., 1994), and in the cortex and hippocampus, the HF stimulus induced suggests a role for Arc in activity-dependent development. Arc only in the hippocampus. Arc mRNA expression was also examined in PC12 cells. The HF stimulation parameters determined to induce Stimulation with nerve growth factor (NGF) or epidermal Arc are identical to those determined previously to induce growth factor (EGF), but not fibroblast growth factor (FG F), LTP, and in each of our preparations, the HF stimulus resulted in a rapid increase in Arc mRNA (Figure 3C). induced robust synaptic enhancement. Since LTP of this Arc mRNA induction by growth factors was enhanced by synapse involves activation of NMDA-type glutamate re- concomitant treatment with the protein synthesis inhibitor ceptors (Abraham and Mason, 1988), we examined the cycloheximide, indicating that Arc is regulated as an lEG hypothesis that NMDA receptor activation is involved in (Nathans et al., 1991). Arc mRNA was not detected by the induction of Arc mRNA. The LTP stimulus protocol was Northern analysis in 3T3 mouse fibroblasts, either unstim- repeated in animals pretreated with the NMDA receptor ulated or stimulated by serum, platelet-derived growth fac- antagonist MK-801 (1 mg/kg). MK-801 blocked the induc- tor (PDGF), FGF, EGF, or PDGF/EGF (data not shown). tion of both LTP and Arc mRNA assayed 30 min following the HF stimulus (n = 3), indicating the synaptic induction Arc mRNA Is Rapidly Regulated in Brain Neurons of Arc involves activation of the NMDA receptor. by Physiological Synaptic Activity The basal level of Arc mRNA expression in the cortex The rapid induction of Arc mRNA following MECS sug- is easily detected by Northern analysis and in situ hybrid- gested that Arc may be regulated by excitatory synaptic ization, and it appears to be comparable to the basal levels mechanisms. To test this hypothesis, granule cells of the of the most abundant lEGs, such as zif268. To examine adult hippocampus were synaptically stimulated by acti- the hypothesis that basal expression in cortex is regulated vating their major afferent projection from the entorhinal by natural synaptic activity, we monitored the effect of cortex using a chronic, in vivo electrophysiological prepa- acutely interrupting synaptic input to the primary visual ration (McNaughton et al., 1986). The intensity and fre- cortex. Afferent activity from the eye is blocked by intra- quency of the synaptic stimulus can be precisely con- ocular injection of the sodium channel antagonist tetrodo- trolled in this preparation, and it is used extensively to toxin (TTX). Visual pathways in the rat are largely crossed study mechanisms of long-term potentiation (LTP). Stimu- (Zilles et al., 1984) such that following a monocular TTX lation at an intensity sufficient to activate granule cell ac- injection, afferent activity is reduced in the contralateral tion potentials, when administered at 0.1 Hz, did not result visual cortex, whereas input to the visual cortex ipsilateral in increased Arc mRNA expression in l~hese cells (Figure to the ocular TTX injection is largely unaffected (Thurlow 4A). By contrast, stimuli of the same intensity, when admin- and Cooper, 1988). Rats sacrificed either 4 or 18 hr after istered at high frequency (HF; 400 Hz), resulted in a rapid monocular TTX injection (n = 3 each) demonstrated ana- and robust induction of Arc mRNA in granule cells. Arc tomically discrete reductions of Arc mRNA in the deaffer- mRNA was induced in each of six animals sacrificed 30 ented visual cortex (Figure 4B). Reductions were more min after the HF stimulus, as well as in each of an addi- pronounced after 18 hr than after 4 hr. Basal expression tional six animals sacrificed 1,2, or 4 hr after the stimulus in cortex of other lEGs is similarly responsive to visual (n = 2 each). No induction was detected in rats sacrificed deafferentation (Worley et al., 1991; Yamagata et al., 24 hr after the HF stimulus (n = 2), suggesting Arc mRNA 1993, 1994a). These findings suggest that natural synaptic expression had returned to basal levels by this time point. activity continuously drives Arc expression in the adult The time course of Arc mRNA induction following the HF cortex. Neuron 436

Figure 3. RegulationofArcmRNAExpression A B HIPPOCAMPUS CORTEX TISSUE DISTRIBUTION (A) Northern analysis of 10 ~g of total RNA per lane prepared from brain and peripheral tis- sues under basal conditions and after MECS u DEVELOPMENT induction. Hippocampus and cortex samples at different time points (0.5, 1, 2, 4, and 8 hr) following MECS reveal that MECS induces a rapid and robust increase in Arc mRNA within 30 min that persists through 8 hr. Tissue distri- Arc xx ~, bution analysis shows that mRNA expres- sion is enriched in brain relative to peripheral tissues, with a low level of expression detected in thymus. (B) Developmental profile of Arc expression. Northern analysis of 10 p.g total RNA per lane prepared from forebrains of rats at the indi- cated postnatal (P) days. Arc mRNA is first de- tected at P8, with a marked increase in expres- C sion between P10 and P21 and a modest decrease to adult levels thereafter. NGF EGF FGF I I I (C)Arcexpression in PC12 cells following addi- tion of growth factors. Arc mRNA is expressed, basally, at prominent levels in PC12 cells and is rapidly and transiently induced by NGF and EGF but not FGF. Pretreatment with cyclohexi- mide enhances Arc induction by these growth factors. There was 10 I~g of RNA per lane; ar- rowheads indicate 3.2 kb Arc message.

Arc mRNA Is Expressed in the Molecular Layer of appeared to be restricted to the cell body layer. For com- the Hippocampus, Suggesting Localization in parison, we demonstrated the in situ hybridization with Neuronal Dendrites probe specific for the enzyme cyclooxygenase type 2 In situ studies provided evidence of an unusual cellular (Cox2; Figure 5B). Cox2 is an lEG that is strongly induced localization of Arc mRNA. Whereas the majority of the in granule cells by MECS, and its mRNA induction exhibits hybridization signal is associated with the cell bodies of the a time course that is essentially identical to that of Arc granule cells in the hippocampus, hybridization extends (Yamagata et al., 1993). Despite a very high level of induc- beyond the cell body layer into the molecular layer, which tion, Cox2 mRNA is exclusively associated with the cell contains dendrites of granule cells (Figure 5A). Expression body layer. in the molecular layer was evident by 1 hr after MECS, Examination of photographic emulsion autoradiograms whereas at earlier time points (15 or 30 min), Arc mRNA at higher magnification demonstrates that the Arc in situ

Figure 4. Arc mRNA Expression Is Regulated by Physiological Activity in the Hippocampus and Cortex (A) In situ hybridization analysis of Arc mRNA in LTP paradigm. Chronically implanted, awake behaving animals received a high frequency (HF) synaptic stimulus to the left dentate gyrus and an identical number of stimuli at low fre- quency (LF) to the right. The HF stimulation of the perforant path stimulus induced LTP. Ani- mals were sacrificed 30 min after the HF stimu- lus. Arc mRNA is strongly induced by the HF stimulus. The lower brain composite image (in- verted) includes half from native control (c) and MECS stimulated (s) rats. (B) Regulation of Arc mRNA expression by nat- ural activity. P21 rats received a unilateral in- travitreal injection of TTX and were sacrificed after 18 hr. This manipulation produced a marked decrease in Arc rnRNA expression in the deprived visual cortex (between arrows). The upper brain image is of an untreated, age- matched control animal. Arc Expression and Regulation in Brain 437

Figure 5. Arc mRNA Is Expressed in the Molecular Layer of the Hippocampus,Suggesting Dendritic Localization In situ hybridizationstudies of control and MECS-stimulatedbrain. (A) Arc mRNA is induced 4 hr after MECS (left half brain) in the granule cell and molecular layers. (B) COX-2 mRNA, assayed in the same animals, is restricted to the granule cell layer. Magnification,7.5 x. (C and D) Bright (C) and dark (D) field microgramsof photographicemulsion from (A). Hybridizationis uniformlydistributed throughout the molecular layer and is not clustered over intrinsic neuronal or glial elements (asterisks)in the molecular layer (m). The border between dentate gyrus and CA1 sector is labeled by arrowheads.Magnification, 250 x.

hybridization signal in the molecular layer is uniformly dis- buffer to isolate crude nuclear and soluble fractions. The tributed and is not associated with foci of cell nuclei (Fig- pellet was sequentially washed with 2 M KCI to solubilize ures 5C and 5D). These data suggest that hybridization membrane-associated proteins and with 1% Triton X-100 in the molecular layer is associated with granule cell den- to solubilize integral membrane proteins. Comparison of drites, as opposed to glial cells or neurons within the mo- the relative amounts of Arc protein in these fractions dem- lecular layer. onstrated that Arc is markedly enriched in the final insolu- ble pellet (Figure 6B). The amount of protein in this final insoluble fraction represents - 75% of the total hippocam- Arc Protein Is Rapidly Induced in Brain and Is pal protein. In other experiments, we monitored Arc in Enriched in a Salt and Triton Insoluble Fraction fractions from cortex prepared using a protocol to purify Arc selective antiserum was generated against a bacterial cytoskeletal proteins (Pardee and Spudich, 1982). Arc was fusion protein and used for Western analysis. Western not solubilized by sequential washes with 0.1 M KCI, 0.05 analysis of hippocampus identified a single band with an M NaHCO3, 1 mM EDTA, or H20 and was enriched in the apparent molecular weight of 55 kDa that was strongly subsequent acetone-washed precipitate. These observa- induced 4 hr after MECS (Figure 6A). The size of the band tions stand in contrast to the predicted solubility of the detected in hippocampus is very similar to that of the TNT protein based on sequence analysis and the observed sol- product and the predicted molecular weight based on the ubility of both the bacterial fusion protein (both the glutathi- cDNA sequence. Based on these observations, we infer one S-transferase and hexahistadine fusion proteins are that the 55 kDa band represents natural Arc. soluble at concentrations of 10 mg/ml) and the in vitro TNT Western analysis was used to examine the solubility protein. The relative insolubility of natural Arc may result properties of natural Arc. Hippocampus was harvested 4 from either posttranslational modifications or interactions hr after MECS and homogenized in an isotonic sucrose with other proteins (see below). Neuron 438

A B course of Arc mRNA following MECS, with detectable in- ~J creases after 1 hr, maximal induction after 3-4 hr, and return to basal levels by 24 hr (data not shown). Higher resolution cellular Iocalizations were obtained using a fluorescent secondary antibody and confocal mi- croscopy in control tissue. Serial optical sections through the neuronal cell soma demonstrate immunoreactivity re- 68=" 68="- stricted to the subplasmalemmal cortex of the soma and dendrite (Figure 8). At the base of the apical dendrite, 43="- 43=" immunostaining is detected as a layer that is contiguous with staining in the soma and extends up the perimeter, but not the center, of the dendrite. This pattern of staining is similar to that of spectrin (Riederer et al., 1986; Zagon et al., 1986) and F-actin (Matus et al., 1982; Caceres and Figure 6. Arc Protein is Rapidly Induced in Brain and Is Enriched in Steward, 1983). a Salt and Triton Insoluble Fraction Immunoblot analysis of Arc. (A) Arc protein is present in control cortex and is induced 4 hr after Arc Cosediments with F-Actin MECS. Based on its homology to spectrin, the solubility properties (B) Arc is enriched in an insoluble cellular fraction that is resistant to of natural Arc, and its localization in the subplasmalemmal 2 M KCI and 1% Triton washes. Tissue fraction preparation is de- region of neurons, we examined the hypothesis that Arc scribed in text. Technical controls for the fractionation protocol include Western blots with antisera specific for c-Jun (nuclear fraction), synap- interacts with cytoskeletal proteins. Full-length Arc was tophysin (membrane associated protein), and (~-spectrin (cytoskeletal prepared in vitro (TNT; Promega) and mixed with actin protein that resists solubilization by 2 M KCI and 1% Triton). under conditions that favor either G-actin (low salt) or F-actin (100 mM KCI; Pardee and Spudich, 1982), and macroaggregates were isolated by centrifugation. Arc Arc ImmunoreacUvity Is Associated with Neuronal cosediments with F-actin (Figure 9A). In control experi- Perikarya and Dendrites ments, we also examined the cosedimentation properties Arc selective antiserum was used for immunohistochemi- of c-Fos, Jun D, Zif268, and murine macrophage migration cal studies of rat forebrain to assess the cellular localiza- inhibitory factor (Weiser et al., 1988; also termed DER-6 in tion and regulation of Arc protein (Figure 7A). In control Lanahan et al., 1992). None of these proteins differentially brain, Arc immunostaining was detected in discrete popu- cosedimented with F-actin, confirming that the Arc interac- lations of neurons in the hippocampus, amygdala, hypo- tion with actin macroaggregates is selective. thalamus, striatum, and cortex, with intense staining of cell Crude actin preparations, such as those used in the bodies and dendritic processes, including distal branches. experiments described above, contain a number of actin- In neocortex, neurons in layers II and VI were most in- associated proteins that cosediment with actin under salt tensely labeled. No immunoreactivity was detected in re- conditions that favor F-actin (Pardee and Spudich, 1982). gions of axons or terminals. The general anatomic pattern Contaminating proteins include tropomyosin and spectrin, of immunostaining parallels the mRNA localization. Within which are part of the cytoskeletal network. To determine a particular structure, the typical pattern of immunostain- if Arc interacts directly with actin or indirectlyvia a contami- ing is intense labeling of a minority of cells. For example, nating protein, we examined Arc cosedimentation using in the dentate gyrus of the hippocampus, we estimate that a more highly purified actin preparation (kindly provided 1%-2% of the granule cells demonstrate intense immuno- by Dr. Thomas Pollard). In these assays, Arc failed to reactivity. Other lEGs demonstrate a similar pattern of dis- cosediment with actin (data not shown). To determine crete cellular staining (Worley et al., 1991; Gass et al., whether the more purified actin might contain a factor that 1993), and in view of their dynamic regulation by natural inhibits Arc cosedimentation, we examined Arc cosedi- synaptic activity, the basal expression pattern may be rep- mentation with mixtures of crude and purified actin prepa- resentative of the activity history of these neurons (Worley rations. In these experiments, Arc cosedimented with an et al., 1991). efficacy that was proportional to the amount of crude actin Following MECS, immunoreactivity increases markedly in the mixture. These studies suggest that Arc cosedimen- in the hippocampus, amygdala, and hypothalamus (Fig- tation involves interactions with an actin-associated protein. ure 7). In the hippocampus, intense immunoreactivity is We also examined whether Arc interacts with crude associated with virtually all of the granule cells. Consistent membranes prepared from hippocampus. Arc cosedimen- with the dendritic localization in granule cells of control tation with the brain membrane preparation was enhanced hippocampus, after MECS, immunoreactivity is detected by salt conditions that favor actin polymerization (Figure throughout the molecular layer and stands in contrast to 9B). the contiguous unstained regions of CA1. In the hypothala- mus, there is also a marked increase in the number of Arc Is a Single Copy Gene, and Homologous Genes intensely immunoreactive cells with prominent localization Are Present in Mouse and Human Genomes in processes. The time course of the induction of Arc im- Genomic Southerns were performed to determine whether munoreactivity in the dentate gyrus parallels the time Arc is a single-copy gene or a member of a gene family. Arc Expressionand Regulationin Brain 439

Figure 7. Arc ImmunoreactivityIs Localizedto Neuronal Soma and Dendrites Hippocampus and hypothalamus of control and MECS-stimulatedrat brain (sacrificed4 hr after MECS). (A) In control hippocempus,a small numberof granule cells show immunoreactivecell bodies and dendrites. (B) After MECS, virtually every granule cell soma displays intense immunostaining, and there is e diffuse enhancementof immunoreac- tivity in the molecularlayer. (C and D) High power views of control and MECS-stimulatedgranule cells. Arc immunore- activityis present in granulecell dendrites.Indi- vidual immunoreactivedendrites are bestvisu- alized in control hippocampus,where positive cells stand out against the low background. (E) Control hypothalamic premammillarynu- cleus displayingfew immunoreactiveneurons. (F) After MECS, there is a dramaticincrease in Arc immunoreactiveneurons, with prominent staining of dendritic processes. Magnification, 20 x (A and B); 400 x (C, D, E, and F).

A single, strongly hybridizing band was detected in EcoRI, (Copeland and Jenkins, 1991). C57BL/6J and M. spretus BamHI, and Hindlll digests of rat genomic DNA hybridized DNAs were digested with several enzymes and analyzed under conditions of either reduced or high stringency. The by Southern blot hybridization for informative restriction rat cDNA probe detects a single, strongly hybridizing frag- fragment length polymorphisms (RFLPs) using a rat cDNA ment in mouse genomic DNA and a single, more weakly Arc probe. The 11.5 M. spretus EcoRI RFLP (see Experi- hybridizing band in human DNA (data not shown). This mental Procedures) was used to follow the segregation indicates that the Arc gene is conserved across species of the Arc locus in backcross mice. The mapping results and suggests an essential role for the gene product. indicated that Arc is located in the distal region of mouse chromosome 15 linked to , Tgn, and Pdgfb. Although Arc Maps to a Region on Chromosome 15 139 mice were analyzed for every marker, up to 158 mice that Contains the Recessive Mouse were typed for some pairs of markers. Each locus was Mutation stargazer analyzed in pairwise combinations for recombination fre- The mouse chromosomal location of Arc was determined quencies using the additional data. The ratios of the total by interspecific backcross analysis, using progeny derived number of mice exhibiting recombinant chromosomes to from matings of ([C57BL/6J x Mus spretus]F1 × C57BL/ the total number of mice analyzed for each pair of loci and 6J) mice. This interspecific backcross-mapping panel has the most likely gene order are as follows: centromere- been typed for over 1600 loci that are well distributed Myc (6/151)-Tgn (9/148)-Arc (3/158)-Pdgfb. The recom bi- among all the autosomes, as well as the X chr()mosome nation frequencies (expressed as genetic distances in cen- Neuron 440

Figure 8. Arc Immunoreactivity Is Localized in the Subplasmalemmal Cortex of the Cell Soma and Dendrites Confocal laser microscopy of Arc immunoftuorescence in a hippocampal granule cell from a control rat. (a-d) Serial 2 pm thick optical sections, in the xy plane, demonstrating prominent immunoreactivity in the periphery of the soma and apical dendrite base. Magnification, 500 x. (e and f) An oblique (xz plane) optical section of the same cell through the base of the apical dendrite (line of section indicated by arrowheads in [e]; magnification, 3000 x ) shows (f) an ovoid region with reduced fluorescence surrounded by a cylinder of intense staining (between arrowheads; magnification, 9000 x ). This unstained area is continuous with a similar cleft at the center of the dendritic shaft seen in (b) and (c). F-actin was localized in adjacent tissue sections using phalloidin-Texas Red and demonstrated a distribution very similar to Arc.

timorgans _ SE) are as follows: Myc (-4.0 _+ 1.6)-Tgn Discussion (-6.1 _ 2.0)-Arc(-1.9 +_ 1.1)-Pdgfb. We have compared our interspecific map of chromo- Arc is a novel lEG that is enriched in brain and is rapidly some 15 with a composite mouse linkage map that reports regulated by neuronal activity. Arc mRNA is strongly in- the map location of many uncloned mouse mutations duced in granule cell neurons of the hippocampus by non- (compiled by M. T. Davisson, T. H. Roderick, A.L. Hillyard, epileptic, NMDA receptor-dependent synaptic stimuli in and D. P. Doolittle and provided from GBASE, a computer- association with physiological synaptic enhancement ized database maintained at The Jackson Laboratory, Bar (LTP). Basal expression in the adult cortex is regulated Harbor, ME). Arc mapped in a region of the composite by natural synaptic activity, as evidenced by rapid reduc- map that contains a mouse mutation with a phenotype tions in Arc mRNA after either interruption of afferent vi- that might be expected for an alteration in this locus (data sual activity or administration of an NMDA receptor antag- not shown). This recessive mutation is stargazer (Stg). onist. In the developing brain, basal expression of Arc Mice homozygous for Stg display spontaneous epileptic mRNA increases markedly during the second and third seizures and frequently tip their heads back and stare postnatal weeks and parallels the period of enhanced upward (Noebels et al., 1990). NMDA-dependent responses of cortical neurons (Williams The distal region of mouse chromosome 15 shares a et al., 1993; Fox and Zahs, 1994), as well as morphologi- region of homology with human chromosomes 8 and 22. cally defined activity-dependent restructuring of synapses In particular, Tgn has been placed on human 8q24 and (Dobbing and Smart, 1974; Greenough et al., 1985). These Pdgfb on 22q 12.3-q13.1. Arc resides between these data suggest a role for the Arc in developmental and adult genes in mouse, suggesting that the human homolog will plasticity. reside on one of these two chromosomes. Arc mRNA appears to be present in dendrites of hippo- Arc Expressionand Regulationin Brain 441

A B Figure 9. Arc Cosedimentswith F-Actin Homogenate Arc was preparedby in vitro transcription/trans- No Actin lmg/ml Actin OmM KCI IOOmMKCI 0tom KCI 100raM KCI 0toNI KCI 100ram KCI lation and its solubilitywas analyzed. S P S P S P S P S P S P (A) Arc was added to low salt (0 mM KCI) or high salt (100 mM KCI, 1 mM MgCI2)buffer and 68~ 68,,.- incubated in the absenceor presenceof actin as describedin the text. Soluble(S) and pellet 43,,"- 43.- (P) fractions were analyzedby SDS-PAGEfol- lowing centrifugation.In the absenceof actin, Arc is solublein both low and high salt condi- tions. In the presenceof actin, Arc remainssol- uble in low salt conditionsbut is concentratedin the pelletin highsalt conditions, which promote actin polymerization.The control, 13 kDa mac- rophagemigration inhibitory factor protein, re- mains in the eupernatantunder all conditions. 14,-- i (B) Arc was addedto hippocampalmembranes in low salt or high salt buffer, and solubleand pellet fractions were analyzedas in (A). The proportion of Arc in the pellet is increased in high salt conditions.

campal granule cell neurons. Recent studies indicate that is not solubilized by high salt (2 M KCI) or nonionic deter- a select group of mRNAs are present in dendrites and gent (1% Triton). The relative insolubility of native Arc, are presumed to be specifically targeted to this structure together with our biochemical studies indicating that Arc (Steward and Banker, 1992; Wilhelm and Vale, 1993). coprecipitates with F-actin, suggests that Arc may be mRNAs present in dendrites include Map2 (Garner et al., linked to macroaggregates of the cytoskeleton in vivo. In 1988) and the a subunit of calcium/calmodulin-dependent preliminary studies, we have used the yeast two-hybrid protein kinase II (Burgin et al., 1990). mRNA targeting has strategy (Chevray and Nathans, 1992) to identify proteins been demonstrated to be important in the development of that interact with Arc. Three of five interacting clones iden- Xenopus oocytes (Melton, 1987) and Drosophila embryos tified thus far correspond to a protein that is homologous (MacDonald and Struhl, 1988; Gavis and Lehman, 1992) to restin, a structural protein from Reed-Sternberg cells and may also be important for the neuronal synapse (Stew- (Bilbe et al., 1992). We are currently investigating the phys- ard and Banker, 1992; Wilhelm and Vale, 1993). Arc is iological significance of this protein interaction. unique among this group of targeted mRNAs in that it is Arc is expressed in discrete populations of neurons of an lEG. the cortex and hippocampus and is enriched in the cell The carboxyl-terminal half of Arc possesses a modest soma and dendrites. Confocal microscopic images indi- degree of homology with the cytoskeletal protein a-spectrin. cate that Arc is present in the cortex of the cell body and This amino acid homology led us to examine the ability dendrites and corresponds to the cellular localization of of Arc to interact with cytoskeletal proteins. Arc, prepared F-actin (Matus et al., 1982; Caceres and Steward, 1983) by in vitro transcription and translation, cosediments with and spectrin (Riederer et al., 1986; Zagon et al., 1986). actin in salt conditions that favor the formation of polymer- Immunohistochemical Iocalizations are therefore consis- ized F-actin. Arc also cosediments with washed brain tent with the hypothesis that Arc interacts with the neu- membranes under identical conditions. Cosedimentation ronal cytoskeleton. with the polymerized actin aggregate is selective for Arc, Structural changes in the dendrite have been described since several other proteins, including members of the in models of neuronal plasticity (Bailey and Kandel, 1993). Fos/Jun family, which possess coiled-coil motifs (Cohen Baudry and Lynch (1988) suggested that influx of calcium and Parry, 1994), do not cosediment in these assays. might activate proteases and cleave spectrin, thereby fa- Cosedimentation of Arc with actin was most robust in cilitating structural rearrangement (Baudry et al., 1988). assays using relatively crude actin, suggesting that Arc Kandel and coworkers have demonstrated transmitter- does not interact directly with actin but rather with an actin- induced structural changes in invertebrate synapses that associated protein. are associated with long-term facilitation or depression Arc is predicted to be a soluble, cytosolic protein based and that are dependent on rapid RNA and protein synthe- on its overall hydrophilic amino acid sequence, absence sis (Schacher et al., 1993). lEGs, such as Arc, that are of a signal sequence or large hydrophobic domains, and targeted to the dendrite might play a direct role in this absence of consensus sequences for glycosylation. Con- activity-dependent response by interacting with and modi- sistent with this analysis, the size of the native protein in fying the function of existing structural proteins. Much as brain is identical to the protein prepared in vitro, indicating growth factor signaling alters cell motility by regulating the that the mature protein is unlikely to be extensively modi- activity of proteins that change the gel-solid state of the fied. However, the majority of native Arc in hippocampus cytoskeleton (Condeelis, 1993; Stossel, 1993; Janmey, Neuron 442

1994), proteins that are rapidly induced could transiently nant protein to Affi-Gel 15 (BioRad) and antibodies eluted in glycine modify the fluidity of the cytoskeleton and effect structural buffer (pH 2.8). changes that persist despite rapid turnover of the lEG. Immunoblotting The many interesting properties of Arc suggest that under- Tissue samples were sonicated in 5 v of 2% SDS loading buffered, standing its regulation and cellular function will provide resolved by SDS-PAGE, and transferred to nitrocellulose. Blots were important insights into neuroplasticity. incubated with the affinity purified antisera (dilution, 1:200) overnight at 4°C and developed using ECL (Amersham). Experimental Procedures Immunocytochemistry Animals and Supplies Experimental animals and age- and sex-matched controls were per- Adult male rats (Sprague-Dawley or Fischer-344) were used. pS]UTP fused with 4% paraformaldehyde in 0.1 M phosphate buffer solution was obtained from NEN-Dupont. MK-801 was obtained from RBI. Kits (PBS). Brains were postfixed for 4-16 hr in the same fixative and for avidin-biotin and immunofluorescence methods were from Vector. subsequently cryoprotected in either 20% glycerol or 30% sucrose. Texas Red-labeled phalloidin and control sera were purchased from Frozen sections, ranging from 25 to 40 #m, were cut on a sliding Sigma. Glial-fibrillary-acidic protein antisera was obtained from Dako. microtome, transferred to PBS, and processed for immunostaining. All other reagents were from Fisher and Sigma. In all instances, sections were preincubated in normal goat serum as blocking agent and then incubated in primary antisera, at dilutions Construction and Screening of the Subtracted cDNA Library between 1:50 and 1:500 for 24-48 hr. Tissue and ceil population local- A subtracted cDNA library was constructed as described (Yamagata izations were detected using the avidin-biotin procedure with horserad- et al., 1993). The subtracted library was screened with [32P]cDNA pre- ish peroxidase detection (Sternberger and Sternberger, 1986). For pared by reverse transcription of poly(A) ÷ RNA prepared from hippo- subcellular Iocalizations, tissue was developed by immunofluores- campus of control or seizure stimulated rats pretreated with cyclohexi- cence and, after epifluorescence microscopic survey, analyzed by con- mide (20 mg/kg, i.p.) as described previously (Lanahan et al., 1992; focal laser scanning microscopy according to a modified protocol (Vin- Yamagata et at., 1993). A near full-length cDNA of rat Arc was isolated cent et al., 1991). These studies included single- and double-labeling by iterative screening of an unsubtracted, oligo(dT)-primed cDNA li- comparisons with GFAP, the transcription factor zif268, and phalloidin brary prepared from hippocampus harvested 4 hr after high-frequency (a toxin that labels F-actin; Wulf et al., 1979). Optical sections were stimulation in rats pretreated with cycloheximide. registered and photographed at 1-5 I~m intervals in the xy and xz planes from control and MECS-stimulated neocortex and hippo- DNA Sequencing campus. Both strands of the full-length Arc cDNA were sequenced as double- stranded plasmids with synthetic, specific primers by the dideoxy- Cellular Fractionation nucleotide chain termination method using deoxyadenosine 5'-[~-3sS] Cortex was harvested from an adult rat 4 hr after MECS treatment thio triphosphate and Sequenase (USB). (see Electrophysiology below). Tissue was homogenized in 10 mM Tris-HCI, 0.5 mM MgCI, and 0.32 M sucrose and was centrifuged at 900 x g to pellet the crude nuclear fraction. Supernatant was removed Northern Analysis and was centrifuged for 100,000 x g/hr. The cytosolic fraction was This procedure was performed as described (Linzer and Nathans, removed, and the pellet was resuspended in a high salt solution of 1983) with 10 ~g of total RNA per lane. The probe used for Northern 2 M KCI, 25 mM Tris-HCI, 1 mM EGTA, and centrifuged for 100,000 x analysis was a 1.0 kb 3'end fragment of Arc cDNA. The cDNA fragment g/hr. The 2 M KCI soluble fraction was removed, and the pellet was was labeled by the random priming technique (Pharmacia) using resuspended in 1% Triton solution and centrifuged for 100,000 x [a-32P] dCTP. g/hr. The Triton fraction and the insoluble pellet were separated, and all fractions were diluted in 2% SDS loading buffer; proportional Genomic Southern amounts of each fraction were analyzed as described in Immuno- EcoRI, BamHI, and Hindlll digested human, mouse, and rat genomic blotting. DNA (10 p.g) each were fractionated by agarose gel, transferred to nitrocellulose in 20 x SSC and probed with a 3.0 kb Arc cDNA labeled Cosedimentation Assays by the random priming technique (Pharmacia) using [(z-32P] dCTP. For actin cosedimentation, rabbit skeletal muscle actin (Sigma) or Prehybridization and hybridization incubations were done in 5 x SSC highly purified actin (from laboratory of Dr. Thomas Pollard, prepared at 65°C. The final wash was done at either high stringency (0.5 x SSC according to Pardee and Spudich, 1982) were resuspended in a low at 68°C) or low stringency (2x SSC at 60°C). salt buffer (10 mM Tris-HCI [pH 8.0], 0.5 mM dithiotreitol, 0.2 mM ATP, 0.2 mM CaCI2). A high salt buffer additionally contained 100 mM KCI In Situ Hybridization and tmM MgCI2. In conditions without actin, 1 p,I of [35S]Arc and DER-6 Freshly dissected brain tissue was rapidly frozen in plastic molds TNT products (Promega) were diluted in either low or high salt buffer. placed on a dry ice/ethanol slurry as described previously (Cole et al., In actin conditions, 1 Id of Arc and DER-6 TNT products were added 1990). Control and experimental tissues were frozen in the same tissue to resuspended actin, for a final volume of 100 Id and a final actin block to assure identical conditions during tissue sectioning, subse- concentration of 1 mg/ml. Mixtures were incubated for 1 hr at 25°C quent storage, and in situ hybridization. ~S-labeled Arc antisense ribo- and then centrifuged for 30 rain at 4°C and 300,000 x g. Supernatants probe was prepared from an appropriately restricted pBS plasmid con- and pellets were recovered and run on t 2.5% polyacrylamide gel and taining the full-length cDNA. In situ hybridization was performed as visualized by autoradiography. Additional samples of supernatants described (Saffen et al., 1988). Selected slides were treated with photo- and pellets were run on 12.5% polyacrylamide gels and visualized by graphic emulsion (Kodak NTB2) and counterstained as previously de- Coomasie staining to monitor the behavior of actin. scribed (Jordan, 1990). For sedimentation with brain homogenate, a rat hippocampus was sonicated in 1 ml low salt buffer. The insoluble fraction was isolated Generation of Polyclonal Antisera by a 5 min microcentrifugation (12,000 RPM). The pellets were resus- Recombinant Arc (amino acids 132-396) was expressed as a bacterial pended in low and high salt buffers, and 1 I11 Arc TNT product was fusion protein using the pTrcHis vector (Invitrogen) according to the added to 99 ~1 of each resuspended pellet. The mixtures were incu- protocol of the manufacturer. The fusion protein was soluble in nonde- bated and centrifuged as above. naturing buffer and was purified over NF+-nitrilotriacetic acid-agarose (Qiagen). Rabbits were immunized with the recombinant Arc protein Electrophysiology using a standard immunization protocol (Cooper, 1994). Antisera was MECS were induced in adult rats using a constant current signal gener- purified over an affinity column prepared by cross-linking the recombi- ator (ECT unit, Ugo Basil) as described previously (Cole et al., 1990). Arc Expression and Regulation in Brain 443

For LIP studies, Fischer-344 rats were implanted bilaterally with stimu- Received September 2, 1994; revised October 12, 1994. lating and recording electrodes in the perforant path and hilus of the dentate gyrus as described previously (Worley et al., 1993). Rats were References allowed to recover for at least 2 weeks before any recordings were performed. Fourteen chronically implanted rats received HF stimula- Abraham, W. C., and Mason, S. E. (1988). Effects of the NMDA recep- tion in one hemisphere and low frequency stimulation in the other tor/channel antagonists CPP and MK-801 on hippocampal field poten- hemisphere. Electrical stimuli consisted of 200 ms diphasic constant tials and long-term potentiation in anesthetized rats. Brain Res. 462, current pulses given at a stimulus intensity of 500 mA. The low fre- 40-46. quency test stimulation was delivered at 0.1 Hz, and the HF stimulation Abraham, W. C., Dragunow, M., and Tate, W. P. (1992). The role of parameters consisted of 50 repetitions of a 20 ms train (i.e., eight immediate early genes in the stabilization of long-term potentiation. pulses) delivered at 400 Hz (400 total pulses). The HF parameters Mol. Neurobiol. 5, 297-314. reliably induce LTP (Dragunow et al., 1989; Jeffery et al., 1990; Abra- Agranoff, B. (1981). Learning and memory: biochemical approaches. ham et al., 1992; Worley et al., 1993). Following this treatment, the In Basic Neurochemistry, Third Edition, G. Siegel, R. Albers, B. Agra- rats were sacrificed at either 30 min (n = 6), 1 hr (n = 2), 2 hr (n = noff, and R. Katzman, eds. (Boston: Little Brown), pp. 801-821. 2), 4 hr (n = 2), or 24 hr (n = 2). Three additional animals were pretreated with MK-801 (1 mg/kg, Alberini, C. M., Ghirardi, M., Metz, R., and Kandel, E. R. (1994). C/EBP is an immediate-early gene required for the consolidation of long-term i.p.) 1 hr prior to the delivery of the HF stimulus. The intensity of the stimulus was adjusted such that the postsynaptic population response facilitation in aplysia. Cell 76, 1099-1114. was identical to that prior to MK-801 administration. Animals were Bailey, C. H., and Kandel, E. R. (1993). Structural changes accompa- sacrificed 30 min after the HF stimulus and processed for in situ hybrid- nying memory storage. Annu. Rev. Physiol. 55, 397-426. ization. Baudry, M., Larson, J., and Lynch, G. (1988). Long-term changes in synaptic efficacy: potential mechanisms and implications. In Long- Monocular Deprivation Term Potentiation: From Biophysics to Behavior. (New York: Alan Monocular deprivation was performed in adult Sprague-Dawley rats Liss), pp. 109-136. as described previously (Worley et al., 1991) using an intravitreal injec- Bilbe, G., Delabie, J., Bruggen, J., Richener, H., Asselbergs, K., Cer- tion of the sodium channel antagonist TTX (10 p.I of 200 p.M solution letti, N., Sorg, C., Odink, K., Wiesendanger, W., DeWolf-Peeters, C., of TTX in PBS). and Shipman, R. (1992). Restin: a novel intermediate filament- associated protein highly expressed in the Reed-Sternberg cells of Interspecific Mouse Backcross-Mapping Hodgkin's disease. EMBO J. 11, 2103-2113. Interspecfic backcross progeny were generated by mating (C57BLJ Brannan, C. I., Gilbert, D. J., Ceci, J. D., Matsuda, Y., Chapman, V. M., 6J x M. spretus)F1 females and C57BL/6J males as described (Cope- Mercer, J. A., Eisen, H., Johnston, L. A., Copeland, N. G., and Jenkins, land and Jenkins, 1991). A total of 205 N2 mice were used to map N. A. (1992). An interspecific linkage map of mouse chromosome 15 the Arc locus (see text for details). DNA isolation, restriction enzyme positioned with respect to the centromere. Genomics 13, 1075-1081. digestion, agarose gel electrophoresis, Southern blot transfer, and hybridization were performed essentially as described (Jenkins et al., Burgin, K. E., Waxham, M. N., Rickling, S., Westgate, S. A., Mobley, 1982). All blots were prepared with Hybond-N ÷ nylon membrane (Am- W. C., and Kelly, P. T. (1990). In situ hybridization histochemistry of ersham). The probe, an - 1.8 kb EcoRI fragment of rat cDNA, was Ca/calmodulin dependent protein kinase in developing rat brain. J. labeled with [~.32p] dCTP using a nick translation labeling kit (Boeh- Neurosci. 10, 1788-1789. ringer Mannheim); washing was done to a final stringency of 0.8 x Caceres, A., and Steward, O. (1983). Dendritic reorganization in the SSCP, 0.1% SDS at 65°C. A fragment of 15.0 kb was detected in denervated dentate gyrus of the rat following entorhinal cortical le- EcoRI-digested C57BL/6J DNA, and a fragment of 11.5 kb was de- sions: a Golgi and electron microscopic analysis. J. Comp. Neurol. tected in EcoRI-digested M. spretus DNA. The presence or absence 214, 387-403. of the 11.5 kb M. spretus-specific EcoRI fragment was followed in Chevray, P. M., and Nathans, D. (1992). Protein-interaction cloning in backcross mice. yeast: identification of mammalian proteins that react with the leucine A description of the probes and RFLPs for the loci linked to Arc zipper of jun. Proc. Natl. Acad. Sci. USA 89, 5789-5792. including myelocytomatosis (Myc), thyroglobulin (Tgn), and Cohen, C., and Parry, D. (1994). s-helical coiled coils: more facts and platelet-derived growth factor beta (Pdgfb) has been reported pre- better predictions. Science 263, 488-489. viously (Brannan et al., 1992). Recombination distances were calcu- Cole, A., Abu-Shakra, S., Saffen, D., Baraban, J., and Worley, P. lated as described (Green, 1981) using the computer program Spretus (1990). Rapid rise in transcription factor messenger RNAs in rat brain Madness. Gene order was determined by minimizing the number of recombination events required to explain the allele distribution pat- after electroshock induced seizures. J. Neurochem. 55, 1920-1927. terns. Condeelis, J. (1993). Life at the leading edge: the formation of cell protrusions. Annu. Rev. Cell Biol. 9, 411-444. Acknowledgments Cooper, H. M. (1994). Preparation of Polyclonal Antisera. (New York: John Wiley and Sons). We thank M. Papapavlou and M. E. Barnsteadfor excellent technical Copeland, N. G., and Jenkins, N. A. (1991). Development and applica- assistance, Bruce Trapp for advice and support in the immunocyto- tions of a molecular genetic linkage map of the mouse genome. Trends chemical analyses, Norman Barker for photographic assistance, and Genet. 7, 113-118. Ratan Bhat, Jennifer Fosnaugh, and Timothy Murphy for help in stud- Dobbing, J., and Smart, J. L. (1974). Vulnerability of developing brain ies of Arc regulation. We also thank Daniel Nathans for help and sup- and behavior. Br. Med. Bull. 30, 164-168. port in the initial cloning of Arc. This work was supported by training Dragunow, M., Abraham, W. C., Goulding, M., Mason, S. E., Robert- grant MH18030 (G. L. L.), and grants EY09374 (P. F. W), AG09219 son, H. A., and Faull, R. L. M. (1989). Long-term potentiation and the (C. A. B. and P. F. W.), H D01046 (W. E. K.), by grants from the Epilepsy induction of c-fos mRNA and proteins in the dentate gyrus of unanes- Foundation of America (W. E. K.), the W. M. Keck Foundation (P. F. W.), thetized rats. Neurosci. Lett. 101, 274-280. the Krieger Mind/Brain Institute (P. F. W.), by a Merck clinician/scientist Flexner, J. B., Flexner, L. B., and Stellar, E. (1963). Memory in mice award (W. E. K.), and by the National Cancer Institute, Department as affected by intracerebral puromycin. Science 141, 57-59. of Health and Human Services, under contract NO1-CO-74101 with A. B. L. Fox, K., and Zahs, K. (1994). Critical period control in sensory cortex. The costs of publication of this article were defrayed in part by Curr. Opin. Neurobiol. 4, 112-119. the payment of page charges. This article must therefore be hereby Garner, C., Tucker, R., and Matus, A. (1988). Selective localization marked "advertisement" in accordance with 18 USC Section 1734 of messenger RNA for cytoskeletal protein MAP 2 in dendrites. Nature solely to indicate this fact. 336, 674-677. Neuron 444

Gass, C., Herdegen, T., Bravo, R., and Kiessling, M. (1993). Induction Nathans, D., Christy, B~ A., Dubois, R., Lanahan, A., Sanders, L. K., and suppression of immediate early genes in specific rat brain regions and Nakabeppu, Y. (1991). Transcription factors induced by growth- by the non-competitive N-methyI-D-aspartate receptor antagonist MK- signaling agents. (Cold Spring Harbor, New York: Cold Spring Harbor 801. Neuroscience 54, 749-758. Laboratory Press). Gavis, E. R., and Lehman, R. (1992). Localization of nanos RNA con- Nedivi, E., Hevroni, D., Naot, D., Israeli, D., and Citri, Y. (1993). Numer- trols embryonic polarity. Cell 71,301-313. ous candidate plasticity-related genes revealed by differential cDNA Goelet, P., Castellucci, V., Schacher, S., and Kandel, E. (1986). The cloning. Nature 363, 713-722. long and the short of long-term memory--a molecular framework. Na- Nguyen, P. V., Abel, T., and Kandel, E. R. (1994). Requirement of a ture 322, 419-422. critical period of transcription for induction of a late phase of LTP. Greenberg, M. E., Ziff, E. B., and Green, L. A. (1986). Stimulation Science 265, 1104-1107. of neuronal acetylcholine receptors induces rapid gene transcription. Noebels, J. L., Qiao, X., Bronson, R. T., Spencer, C., and Davisson, Science 234, 80-83. M. T. (1990). Stargazer: a new neurological mutant on chromosome Greenough, W. T., Hwang, H.-M. F., and Gorman, C. (1985). Evidence 15 in the mouse with prolonged cortical seizures. Epilepsy Res. 7, for active synapse formation or altered postsynaptic metabolism in 129-135. visual cortex of rats reared in complex environments. Proc. Natl. Acad. Pardee, J. D., and Spudich, J. A. (1982). Purification of muscle actin. Sci. USA 82, 4549-4552. Meth. Enzymol. 85, 164-181. Janmey, P. A. (1994). Phosphoinositides and calcium as regulators Qian, Z., Gilbert, M. E., Colicos, M A., Kandel, E. R., and Kuhl, D. of cellular actin assembly and disassembly. Annu. Rev. Physiol. 56, (1993). Tissue-plasminogen activator is induced as an immediate-early 169-191. gene during seizure, kindling and long-term potentiation. Nature 361, Jeffery, K. J., Abraham, W. C., Dragunow, M., and Mason, S. E. (1990). 453-457. Induction of Foe-like immunoreactivity and the maintenance of long- Riederer, B. M., Zagon, I. S., and Goodman, S. R. (1986). Brain spec- term potentiation in the dentate gyrus of unanesthetized rats. MoI. trin (240/235) and brain spectrin (240/235E): two distinct spectrin sub- Brain Res. 8, 267-274. types with different locations within mammalian neural cells. J. Cell Jenkins, N. A., Copeland, N. G., Taylor, B. A., and Lee, B. K. (1982). Biol. 102, 2088-2097. Organization, distribution, and stability of endogenous ecotropic mu- Saffen, D., Cole, A., Worley, P., Christy, B., Ryder, K., and Baraban, J. rine leukemia virus DNA sequences in chromosomes of Musmusculus. (1988). Convulsant-induced increase in transcription factor messenger J. Virol. 43, 26-36. RNAs in rat brain. Proc. Natl. Acad. Sci. USA 85, 7795-7799. Jordan, C. A. (1990). In situ hybridization in cells and tissue sections: Schacher, S., Kandel, E. R., and Montarolo, P. G. (1993). cAMP and a study of myelin during CNS myelination and remy- arachidonic acid stimulate long-term structural and fu nctional changes elination. In In Situ Hybridization Hietochemistry. (Boca Raton, Florida: produced by neurotransmitters in Aplysia sensory neurons. Neuron CRC Press), pp. 39-70. 10, 1079-1088. Kaufmann, W. E., Yamagata, K., Andreasson, K. I., and Worley, P. F. Shaw, G., and Kamen, R. (1986). A conserved AU sequence from the (1994). Rapid response genes as markers of cellular signaling during 3' untranslated region of GM-CSF mRNA mediates selective mRNA cortical histogenesis: their potential in understanding mental retarda- degredation. Cell 46, 659-667. tion. Int. J. Dev. Neurosci. 12, 263-271. Sheng, M., and Greenberg, M. (1990). The regulation and function of Kozak, M. (1987). At least six nucleotides preceeding the AUG initiator c-foe and other immediate early genes in the nervous system. Neuron codon enhance translation in mammalian cells. J. Mol. Biol. 196, 947- 4, 477-485. 950. Sternberger, L. A., and Sternberger, N. H. (1986). The unlabeled anti- Kozak, M. (1989). The scanning model for translation: an update. J. body method: comparison of peroxidase-antiperoxidase with avidin- Cell Biol. 108, 229-241. biotin complex by a new method of quantitation. Histochem. Cyto- Lanahan, A., Williams, J. B., Sanders, L. K., and Nathans, D. (1992). chem. 34, 599-605. Growth factor-induced delayed early response genes. Mol. Cell. Biol. Steward, O., and Banker, G. A. (1992). Getting the message from the 12, 3919-3929. gene to the synapse: sorting and intracellular transport of RNA in Lau, L F., and Nathans, D. (1991). Genes induced by serum growth neurons. Trends Neurosci. 15, 180-186. factors. In The Hormonal Control of Gene Transcription, Volume 6, Stossel, T. (1993). On the crawling of animal cells. Science 260, 1086- Molecular Aspects. (Amsterdam: Elsevier Science Publishers), pp. 1094. 257-293. Thurlow, G. A., and Cooper, R. M. (1988). Metabolic activity in striate Linzer, D. I. H., and Nathans, D. (1983). Growth-related changes in and extrastriate cortex in the hooded rat: contralateral and ipsilateral specific mRNAs of cultured mouse cells. Proc. Natl. Acad. Sci. USA eye input. J. Comp. Neurol. 274, 595-607. 80, 4271-4275. Vincent, S. L, Sorensen, I., and Benes, F. M. (1991). Localization and MacDonald, P. M., and Struhl, G. (1988). Cis-acting sequences respon- high-resolution imaging of cortical neurotransmitter compartments us- sible for anterior localization of bicoid mRNA in Drosophila embryos. ing confocal laser scanning microscopy: GABA and glutamate interac- Nature 336, 595-598. tions in rat cortex. Biotechniques 11,628-634. Matus, A., Ackermann, M., Pehling, G., Byers, H. R., and Fujiwara, Wasenius, V. M., Saraste, M., Salven, P., Eramaa, M., Holm, L., and K. (1982). High actin concentrations in brain dendritic spines and post- Lehto, V. (1989). Primary structure of the brain (~-epectrin. J. Cell Biol. synaptic densities. Proc. Natl. Acad. Sci. USA 79, 7590-7594. 108, 79-93. McNaughton, B. L., Barnes, C. A., Rao, G., Baldwin, J., and Rasmus- Weiser, W. Y., Temple, P. A., Witek-Giannotti, J. S., Remold, H. G., sen, M. (1986). Long-term enhancement of hippocampal synaptic Clark, S. C., and David, J. R. (1988). Molecular cloning of a cDNA transmission and the acquisition of spatial information. J. Neurosc}. encoding a human macrophage migration inhibitory factor. Proc. Natl. 6, 563-571. Acad. Sci. USA 86, 7522-7526. Melton, D. (1987). Tranelocation of a localized maternal mRNA to the Wilhelm, J. E., and Vale, R. D. (1993). RNA on the move: The mRNA vegetal pole of Xenopus oocytes. Nature 328, 80-82. localization pathway. J. Cell Biol. 123, 269-274. Montarolo, P. G., Goelet, P., Castellucci, V., Morgan, J., Kandel, E., Williams, K., Russell, S. L., Shen, Y. M., and Molinoff, P. B. (1993). and Schacher, S. (1986). A critical period for macromolecular synthe- Developmental switch in the expression of NMDA receptors occurs sis in long-term heterosynaptic facilitation in Aplysia. Science 234, in vivo and in vitro. Neuron 10, 267-278. 1249-1254. Worley, P. F., Cole, A. J., Murphy, T. M., Christy, B. A., Nakabeppu, Moon, R. T., and McMahon, A. P. (1990). Generation of diversity in Y., and Baraban, J. M. (1990). Synaptic regulation of immediate early nonerythroid spectrins. J. Biol. Chem. 265, 4427-4433. genes in brain. Cold Spring Harb. Symp. Quant. Biol. 55, 213-223. Arc Expression and Regulation in Brain 445

Worley, P. F., Cole, A. J., Murphy, T. M., Christy, B. A., Nakabeppu, Y., and Baraban, J. M. (1991). Constitutive expression of zif268 in neocortex is regulated by synaptic activity. Proc. Natl. Acad. Sci. USA 88, 5106-5110. Worley, P. F., Bhat, R. V., Baraban, J. M., Erickson, C. A., McNaugh- ton, B. L., and Barnes, C. A. (1993). Thresholds for synaptic activation of transcription factors in hippocampus: correlation with long-term en- hancement. J. Neurosci. 13, 4776-4786. Wulf, E., Deboben. A., Bautz, A., Faulstich, H., and Wieland, T. (1979). Fluorescent phallotoxin, a tool for the visualization of cellular actin. Proc. Natl. Acad. Sci. USA 76, 4498-4502. Yamagata, K., Andreasson, K. I., Kaufmann, W. E., Barnes, C. A., and Worley, P. F. (1993). Expression of a mitogen-inducible cyclooxy- genase in brain neurons: regulation by synaptic activity and glucocorti- coids. Neuron 11,371-386. Yamagata, K., Kaufmann, W. E., Lanahan, A., Papapavlou, M., Barnes, C. A., and Worley, P. F. (1994a). Egr3/Pilot, a zinc-finger tran- scription factor, is rapidly regulated by activity in brain neurons and co-localizes with Egr1/Zif268. Learn. Mem. 1, 140-152. Yamagata, K., Sanders, L. K., Kaufmann, W. E., Barnes, C. A., Na- thans, D., and Worley, P. F. (1994b). rheb, a growth factor and synaptic activity regulated gene, encodes a novel Ras-related protein. J. Biol. Chem. 269, 16333-16339. Yan, Y., Winograd, E., Viel, A., Cronin, T., Harrison, S. C., and Branton, D. (1993). Crystal structure of the repetitive segments of spectrin. Science 262, 2027-2030. Zagon, I. S., Higbee, R., Riederer, B. M., and Goodman, S. R. (1986). Spectrin subtypes in mammalian brain: an immunoelectron micro- scopic study. J. Neurosci. 6, 2977-2986. Zilles, K., Wree, A., Schleicher, A., and Divac, I. (1984). The monocular and binocular subfields of the rat's primary visual cortex: a quantitative morphological approach. J. Comparative Neurol. 226, 391-402.

GenBank Accesion Number

The GenBank accession number for the sequence reported in this paper is U19866.