Proc. Natl. Acad. Sci. USAVol. 89, pp. 3020-3024, April 1992Neurobiology

Mouse 5HT1B serotonin receptor: Cloning, functional expression,and localization in motor control centers

(cydic AMP/guanine nucleotide-binding proteln/strlatum/cerebellum/Purkinje cells)

L. MAROTEAUX, F. SAUDOU, N. AMLAIKY, U. BOSCHERT, J. L. PLASSAT, AND R. HEN*Laboratoire de G6ndtique Mol6culaire des Eucaryotes du Centre National de la Recherche Scientifique, Unit6 184 de Biologie Mol6culaire et de GenieG6n6tique, Institut National de la Sante et de la Recherche M6dicale, Facult6 de M6decine, 11 rue Humann, 67085 Strasbourg C6dex, France

Communicated by Richard Axel, January 2, 1992

ABSTRACT Serotonin is a neuromodulator that mediatesa wide range of effects by interacting with multiple receptors.Using a strategy based on nucleotide sequence homology be-tween genes encoding receptors that interact with guaninenudeotide-binding proteins, we have isolated a mouse geneencoding an additional serotonin receptor. When expressed incultured cells, it displayed the pharmacological profile andcoupling with adenylate cyclase characteristic of the 5HT1Breceptor subtype. In NIH 3T3 cells expressig this receptor,serotonin induced a decrease in forskolin-stimulated cAMPlevels. This effect was blocked by pertussis toxn, indicatingthat the 5HT1B receptor interacts with a pertussis toxin-sensitive guanine nucleotide-binding protein. To obtain clues asto the possible function of the SHT1B receptor, we haveanalyzed its pattern of expression in the adult mouse brain byin situ hybridization. Our results, together with previousautoradiographic studies, suggest that the 5HT1B receptors arelocalized presynaptically on the terminals of striatal neuronsand Purkinje cells and that they might modulate the release ofneurotransmitters such as y-aminobutyric acid. The predom-inant expression of the 5HT1B receptor in the striatum andcerebellum points to an involvement of this receptor in motorcontrol.

Serotonin (5-HT) is a neuromodulator that is involved invarious functions such as sleep, appetite, pain perception,and vascular contraction. This diversity of effects can berelated to the fact that the 5-HT-ergic neurons project intovirtually all parts of the brain and spinal cord, although theircell bodies are concentrated in a limited area, the raphenuclei. 5-HT activates multiple receptor subtypes that exhibitdistinct pharmacological properties, signaling systems, andtissue distributions (for a review, see ref. 1). The 5HT1Breceptors have been identified in the rat and mouse brainwhere their highest density was found within the globuspallidus and the substantia nigra. However, they could not bedetected in the brain of other species, including humans.These species contained, instead, 5HT1D receptors that havea slightly different pharmacological profile but the sametissue distribution. It was therefore suggested that the 5HT1Band iD receptors correspond to species variants of a samereceptor subtype. The 5HT1B and the 5HT1D receptors arenegatively coupled with adenylate cyclase. Recent cloning ofthe 5HT1A, 5HT1C, and 5HT2 receptors has revealed thatthey belong to the large family of receptors that interact withguanine nucleotide-binding proteins (G proteins) and share apredicted seven-transmembrane-domain structure (2). Wehave exploited the sequence homologies that exist betweenseveral members of this family to clone the gene encoding themouse 5HT1B receptor.t Our results suggest that the 5HT1B

receptors are localized presynaptically on the terminals ofstriatal neurons and Purkinje cells.

MATERIAL AND METHODSIolation and Sequence of the SHT1B Genoinc Clone. A

nested PCR experiment was performed on mouse genomicDNA with the following oligonucleotides: (i) TACCTCGAG-GTCGACGGTITG(C/T)TGG(C/T)TICCITT(C/T)TT; (ii)AGAACTAGTGGTACCC(G/A)TIGT(G/A)TA(G/A/T)ATIA(C/T)IGG(G/A)TT; (iii) AGAACTAGTGGTAC-CC(G/C)(T/A)(G/A)TTIAC(G/A)TAICCIA(A/G)CCA.One microgram of DNA was annealed at 550C and amplifiedat 720C in the presence of 3 mM MgCl2 for 20 cycles withprimers i and ii and for 20 more cycles with primers i and iii.The PCR products were cut with Xho I and Kpn I cloned inthe Bluescript plasmid and sequenced. One of the deducedamino acid sequences resembled that of 5-HT-ergic recep-tors. The corresponding oligonucleotides were synthesized(TGGCCATGTGAAACCAGCAGGCATC; TTCCCTGGT-GATGCCTATCGTAAG) and used to screen a mouse ge-nomic library. The Bgl II-Sac I fragment (Fig. 1) hybridizingwith these two oligonucleotides was sequenced on bothstrands by the dideoxynucleotide technique using successivesynthetic oligonucleotides.

Expression of the 5HT1B Receptor in Cultured Cells. TheBgl II-Sac I genomic fragment (Fig. 1) was inserted into theBgl II and Sac I sites of expression vector p513, which is aderivative of pSG5 (3) containing a multiple cloning site. Theresulting recombinant was introduced into mouse NIH 3T3cells by calcium phosphate-mediated transfection togetherwith the recombinant pRSVneo, which encodes resistance toG418 (20 ,ug of5HT1B recombinant and 1 jig ofpRSVneo per10-cm dish). Transformed clones were selected in the pres-ence of 0.5 mg of G418 per ml. Isolated foci were amplifiedand total RNA was prepared and analyzed for expression of5HT1B mRNA. Two cell lines were selected that expressedhigh levels of 5HT1B mRNA as determined by Northern blotanalysis.For transient expression of the 5HT1B receptor, COS-7

cells were transfected by the calcium phosphate techniquewith the 5HT1B recombinant alone (20 jug per 10-cm dish)and analyzed 48 hr after transfection.

Radioligand Binding Assays. Membranes were prepared (4)and [3H]5-HT binding assays and competition displacementexperiments were performed as described (5).cAMP Assays. Cells were seeded into 12-well plates at a

density of =3 x 105 cells per well, washed once withphosphate-buffered saline (PBS), and incubated for 15 min at

Abbreviations: G protein, guanine nucleotide-binding protein; 5-HT,5-hydroxytryptamine.*To whom reprint requests should be addressed.tThe sequence reported in this paper has been deposited in theGenBank data base (accession no. M85151).

3020

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Neurobiology: Maroteaux et al.

370C with 100 ,uM isobutylmethylxanthine and test agents inPBS. The reaction was stopped by aspiration of the mediumfollowed by addition of 500 ,ul of ice-cold ethanol. After 2 hrat room temperature, the ethanol was collected and Iyophi-lized. The pellet was reconstituted and cAMP was quanti-tated using a radioimmunoassay kit (NEN, NEK-033). Thebasal level of cAMP observed in the absence of drugs wasabout the same in all cell lines (-300 pmol/mg of protein).One micromolar forskolin typically yielded a 10-fold increasein cAMP levels.RNA Analysis. Poly(A)+ mRNA was prepared, fractionated

on a 1% agarose/formaldehyde gel, and transferred to a nitro-cellulose filter. The DNA probe was the Bgl TI-Sac I fragment,which was 32P-labeled by random priming and hybridized tofilters at high stringency [42°C; 50%o formamide/5 x SSC (1 xSSC = 0.15 M NaCl/15 mM sodium citrate)/1 x Denhardt'ssolution/20 mM sodium phosphate buffer, pH 6.5/0.1% SDS/100 ,g of tRNA per ml]. Washings were performed at highstringency (60°C; 0.1 x SSC/0.1% SDS).In Situ Hybridization. In situ hybridizations were per-

formed on cryostat sections of adult mouse brains (about 8

Proc. Natl. Acad. Sci. USA 89 (1992) 3021

weeks old) as described (6). The probe used was the 5HT1BBgl II-Sac I genomic fragment labeled by random primingwith ATP[35S]. Slides were exposed for 10 days.

RESULTSIsolation ofa Mouse Genomic Clone Encoding an Additional

Member oftheG Protein-Coupled Receptor Family. Sequencecomparisons ofG protein-coupled receptors have revealed astriking amino acid sequence conservation, particularly incertain putative transmembrane domains such as domains VIand VII. We decided therefore to use degenerate oligonucle-otides corresponding to these two regions to perform a seriesof PCR experiments on mouse genomic DNA. The resultingfragments were subcloned and sequenced. One of thesefragments was used to screen a mouse genomic library. Weobtained two phage recombinants that contained a 2.3-kilobase-pair-long Bgi II-Sac I fragment (Fig. 1) hybridizingwith the PCR product. Sequence analysis of this fragmentrevealed one long open reading frame encoding a predictedprotein that exhibited highest homology to the human 5HT1Dreceptor (59o) (7) and to the rat 5HT1A receptor (47%) (8).

1 GAG CTC CGG GCT TGT AGC GCC CAT GCC GTG GCC ATG91 TCC TTC GGG CGG AGT CTC CAG ATT CCG CAG CGC CCC

181 GAG ACC CCT CTA CAA GCG ATA CTT ACT CCT TCT CCA271 TTC ACC CCC GCG AAG ACG GCC AGT CGA GAG GAG CTA361 GCT TGG TCC CCC CCC CTG L CCT CTC CTT TGG GCT451 AGA CAA GCC TAT ACT CTC CAT CAT CCT CCC GTC CTC541 AGC AGA CCT ATG GAG GAG CAG CCT ATT CAG TCC GCC

M E E Q G I Q C A

CAA CCC CAG GCAC TCT CCA CTGGCT TAG GAG AAG CTG AGA CCTCTG CTT CCC TCC ACC CAA CCTCCA GAC GCG CCA GCA CGC CCTGAG AAC ACA GCC GGA CGA GCTTCTI TCT TTC CCT GCC CCG CTCCCC CCG CCT CCC GCC CCC TCCP p p P A A S

CCC TGC AAG CTT CGG TCT CCA CAG CCACAG CAA TCG GGG TAA CAG CCC CCA CTAGAG ACA AGC TCC ACT AGG TGG GCC AGTGGC CGA CTG GAA CTG CAG GGG ACG CGCACT GAG GAG CCC ACG GAA CTIG GCT AGCCAT GTT CAA GAG CTG CGC TCC GCA GCCCAG ACA CCG GTA CCT CTC ACC AAC CTCQ T G V P L T N L

I ,W631 AAC TGC AGC GCC GAC GCC TAC ATT TAC CAG GAC TCC ATC CCC CTG CCC TC AA GTC G CTG GTTGA TG TTG GG CTC ATC ACCAG

'mN C S A D G Y I Y Q D S I A L P W K V L L V A L L A L I T L

721 GCC ACC ACG CTC TCC AAC GCC TTT GTA ATC GCT ACG GTG TAT CGG ACC CCC AAC CTA CACA T T L S N A F _V I A T V Y R T R K L H

I I 4)811 GCA GTC ACT GAC CTG CTC GTG TCC ATC CTG GTG ATG CCC ATC ACC ACC ATG TAC ACG TC

A V T D L L V S I L v M P I S T M Y T VI I I

901 TCC CAC TTC TGG CTG TCG TCG CAT ATC ACC TGT TGC ACT CCT TCC ATC ATC CAT CTC TGTC D F W L S S D I T C C T A S I M H L C

991 ACC GAT GCG

T D A

1081 TTG CCA CCC

L P P

1171 TCC ACG GTG

S T V

1261 CAG ACA CCC AAC

Q T P N

1351 TCC CGG CCT CCGS R A P

GTG GAG TAT TCT

v E Y S

O

GCT AAA AGG ACT CCC AAA AGG GCG GCC ATC ATG ATC

A K R T P K R A A I M I

TTC TTC TGG CCT CAA GCC AAA

F F W R Q A K

V

GGC GCT TTC TAT TTA CCC ACC

G A F Y L P T

AAG ACC GCCK T C

GAC GTG CCC

D v P

GCA GAG GAG GAC ATG CTG GAC TGC TTTA E E E M L D C F

CTGV

GTG

ACC CCG CCT AAC TAC CTG ATC CCC TCT CTGT P A N Y L I A S L

ACT GGA CCC TGG ACA CTA GGC CAG GTG GTCT G R W T L G

87

Q V v 117

GTC ATC GCC CTG GAC CGC TAC TGG GCC ATC

V I A L D R Y W A I 147IV

CTG CCG TCC GTC TTC TCC ATC TCT ATT TCGL V W v F S I S I S 177

AAC ACC CGAC CAC GTC CTC TAC ACG GTC TAC

N T D H v L Y T v Y 207

CTG CTC CTC ATC GCC CTC TAT GGC CCC ATC TAT CTG GCAA CC CGC TCT CGG ATT TTG AAAL L L I A L Y G R I Y V E A R S R I L K 237

AAG CCC TTG ACC CGA GCC CAG TTG ATA ACA GAC TCCK R L T R A Q L I T D S

AGT GAG TCC GGG TCT CCT CTG TAC CCG AAC CAA GTCS E S G S P v Y v N Q V

1441 AAG AAA CTC ATG GCC GCT ACG CAG CCC AAA GCC ACC AAG ACC | CTA G ATC ATT TTAK K L M A A R E R K A T K T L G I I L

CCG GCA TCC ACA TCC TCG GTC ACC TCC ATT AACP G S T S S V T S I N 267

AAA GCG CGA GTC TCA GAC CCC CTG CTG GAA AAGK V R v S A L L E K 297

VICCA GCA TTT ATT GTG TGT TCG CTG CCC TTC TTCG A F I V C W L P F F 327

VI T

1531 ATC ATC TCC CTG

I I S L

1621 TCC CTC ATC AACS L I N

GTG ATC1CCT ATCv M

CCC ATC ATC TAC

P I I Y

1711 CTT GTC AAT GGC ACT GGG1801 GCG TGG AAG CAA GGC TCT1891 GCA ATG TCA ACA ACG CAA1981 CAG CAG CCA TTC ACA CAA

2071 CCT CTA ACC CAG TAG CAC

2161 TCT ATC CTC TGG GCA GAA2251 GAG AAA CTT GTG GAT TTC

2341 TAA CAT CT

TGC AAC GAT GCC TCC TGG TTT CAC

C K D A C W F H

ACC ATC TCC AAT CAG CAC TTC AAA CAA CGC TTCT M S N E D F K Q A F

a

CTC CCC TAA CGC ACC TT GGC CAC CAA CTT CCC TCT TCTCTC TTC TCC CCA ACC CCA ATG CAT CCT CAG ACC CGC CCCACT ACT GTA AAG CCT CCA TAC CCG ACC AMC TTA CCT TCCAAG CTT GCA ACT TTT TAA CAT TGG TCA TGG ATG GCT ATC

TTG TAG ATG TTG TCI GAG GCC TGA GAC ACC TCT ACA TACCTC TGT GCT ACA GCT TM GCT CTA MC ACT TTG CCA TCC

GTG CTT CAA ACA CTC AAT GTG GCT TCG ACG CCA CAG AGG

ATG CC ATT CTT GAC TTC TTC AAT TGG T1A GGC TAT CTT AAC.M A I F D F F N W 1, G Y L N 357

CAC AAA CTC ATA CGC TTT AAG TGC GCA GGT TGAH K L I p F K C A G 386

CAC AGC TAG GTC AAA TCT TCT TTCCCA GGA AGC CTA CCT TCT CTG AGATCC ACT CM ACT CCG CGC TCA GCACCT GCC CCG GTT TTG TAT CAT GGAAGC CTG GCA GTA CTT CM CTT GACCTC ACT CTT CAT CTA TTT ATT TAAAAC CCT GAA GAC TA ACA GCA AM

CCT GTT ACTCAG CAC TCTCAG ATC CTC

TGT TGT TGCAAT TAA TACATT TCG TCGTTC TGA TGC

FIG. 1. Nucleotide sequence ofthe Sac I-Bgl II genomic fragment encoding the 5HT1B receptor. The seven putative transmembrane domainsare boxed and numbered (I-VII). Arrows indicate sites of potential N-linked glycosylation. *, A, and *, consensus sites for phosphorylationby protein kinase C, protein kinase A, and tyrosine kinase, respectively.

GCA AACGAC CCCGCT [1ACCC CAGCAA AGGAGG ACGTCC CAC

S H 27

57

3022 Neurobiology: Maroteaux et al.

Hydropathy analysis of this predicted protein revealed sevenhydrophobic domains (numbered I-VII in Fig. 1), a featureshared by all other cloned members of the G protein-coupledreceptor family. The amino-terminal end contained two pu-tative sites for N-linked glycosylation and the presumedcytoplasmic domains contained several consensus sites forphosphorylation by protein kinases C and A (Fig. 1).

The' Mouse Receptor Has the Same Parma logica Profileas the 5HT1B Receptor. To determine whether the genomicfragment that we had isolated encoded a functional receptor,we introduced it into a eukaryotic expression vector andtransfected COS-7 cells with the resulting recombinant.Membranes of transfected cells were then assayed for theirability to bind a number of5-HT-ergic radioligands. Although

I-labeled lysergic acid diethylamide, 3H-labeled ketan-serine, and 3H-labeled 8-hydroxy-2-(di-n-propylamino)tetra-lin ([3HJ8-OH-DPAT) did not bind to these membranes,[3H]5-HT displayed a single saturable binding site: Kd = 48nM and Bma, = 19 pmol/mg of membrane protein (notshown). In a control experiment, [3H]5-HT did not bind tomock-transfected COS-7 cells. To determine the pharmaco-logical profile ofthis receptor, bound [3H]5-HT was displacedwith various 5-HT-ergic drugs. These compounds displayedthe following rank order of potencies: cyanopindolol >5-carboxamidotryptamine-(5-CT) = RU24969 > 5-HT >(-)-pindolol (Table 1). Ketanserine, mianserine, yohimbine,spiperone, and 8-OH-DPAT were almost inactive. This pro-file corresponds well with that of the rat brain 5HT1Breceptor (Table 1).

Table 1. Pharmacological profile of the 5HT1B receptorBinding, pKd' Cyclase, pEC5o or pK1

Mouse Mouse RatCOS-7 Rat NIH 3T3 substantia

Drug cells cortex cells nigraAgonist5-HT 7.4 (4) 7.6 7.8 (3) 7.85-CT 8.0 (3) 8.3 8.1 (2) 7.9RU24969 8.0 (3) 8.4 8.3 (2) 8.4

AntagonistCyanopindolol 8.6 (3) 8.3 8.8 (3) 8.2Methiothepin - 7.3 8.4 (2) 8.1(-)-Pindolol 7.2 (2) 7.2 6.8Mianserine 5.4 (2) 5.3 5.9Ketanserine 5.2 (2)Yohimbine 4.7 (2) 5.5 6.1Spiperone 4.6 (2) 5.3 4.48-OH-DPAT 4.5 (2) 4.2 4.8Binding data correspond to competition for [3H]5-HT binding to

membranes of COS-7 cells expressing the mouse SHT1B receptor.ICso values required to displace 50%1 of [3H15-HT were determinedexperimentally and converted to pKd' values according to theequation Kd' = IC50/(1 + C/Kd), where C is the [3H]5-HT concen-tration (25 nM) and Kd is the equilibrium dissociation constant of[3H]5-HT (48 nM). Cyclase data were obtained with an NIH 3T3-derived cell line expressing the mouse 5HT1B receptor. EC50 is theconcentration of agonist required to obtain a half-maximal inhibitionof forskolins-stimulated adenylate cyclase. The concentrations ofantagonist required to inhibit 5-HT's effect by 50%i0 (IC50) weredetermined and converted to pK, values according to the equation K,= IC50/(1 + C/lKd), where C is the 5-HT concentration (100 nM) andKd is the EC50 value for 5-HT (14 nM). Binding and cyclase data werecompared to those reported in rat cortex and rat substantia nigra (9).Numbers in parentheses correspond to the number of independentexperiments performed. The values presented are the mean of atleast two independent experiments (each determination performed intriplicate). Individual Kd', EC50, and Ki values differed by <20%6.5-CT, 5-carboxamidotryptamine; 8-OH-DPAT, 8-hydroxy-2-(di-n-propylamino)tetralin.

E

co

T

0)

0.U-is0O

100

90-

80-

70

60-

50-

40ft I,

10 9 8]6 5 4

-log [ S-HT ]

FIG. 2. 5-HT induced a decrease in cAMP levels in NIH 3T3 cellsexpressing the 5HT1B receptor. cAMP levels were expressed as apercentage of the value obtained with 1 gM forskolin (100%). Dataare the mean of three independent experiments (each determinationperformed in triplicate). e, 5-HT plus forskolin; e, 5-HT, forskolin,and pertussis toxin. Pertussis toxin was applied at a concentration of100 ng/ml 20 hr before the addition of 5-HT and forskolin.

5-HT Inibits Adenylate Cycase in NIH 3T3 Cells Express-ing the SHT1B Receptor. To analyze the coupling of the5HT1B receptor to the second messenger machinery, wegenerated stable clonal cell lines expressing this receptor. Intwo independent cell lines expressing high levels of 5HT1BmRNA, 5-HT mediated a decrease in forskolin-stimulatedcAMP levels but had no effect on control NIH 3T3 cells. Thisdecrease in cAMP level was concentration dependent andsaturable, the EC50 for5-HT being 1.4 x 10-8M (Fig. 2). Twoother 5HT1B agonists, 5-CT and RU24969, also induced adecrease in cAMP levels, with EC50 values of8 x 10-9M and5 x 10-9 M, respectively (Table 1). The effect of5-HT couldbe blocked by cyanopindolol and methiothepin, which are5HT1B antagonists. The resulting EC50 and Ki values (Table1) are in good agreement with those we obtained in bindingassays and with the values reported in rat substantia nigra forthe 5HT1B receptor (9). Pertussis toxin blocked the effect ofserotonin (Fig. 2), indicating that the 5HT1B receptor iscoupled to a pertussis toxin-sensitive G protein.The 5HT1B Receptor Is Expressed Pr tly in the

Striatum and Cerebellum. Expression of 5HT1B transcriptswas analyzed by Northern blot and in situ hybridizationexperiments. The Northern analysis revealed that expressionwas predominant in nervous tissue. A 6-kb transcript wasdetected in forebrain, hindbrain, cerebellum, and spinal cord(Fig. 3 and results not shown). To further analyze the patternof expression of this receptor we performed in situ hybrid-ization experiments on brain sections (Fig. 4). The main sitesof expression were the caudate-putamen and the Purkinje

-~28 cFIG. 3. Distribution of SHT1B

transcripts. Northern blot analysis ofpoly(A)+ RNA (5 1&g) from variousorgans. A 6-kb RNA was detected incerebellum and total brain and de-

_18a- tected very faintly in kidney. Theprobe used was the 32P-labeled SacI-Bgl II fragment.

Proc. Nad. Acad Sci. USA 89 (1992)

Proc. Natl. Acad. Sci. USA 89 (1992) 3023

FIG. 4. In situ hybridization to mouse brain horizontal sections. (a-c) Dark-field pictures or successively more dorsal sections of a wholebrain (8 mm wide). (d) A 1.5-fold magnification (compared to a-c) of the striatal region at a depth intermediate between that of b and c. (e) A4-fold magnification (compared to a-c) ofa cerebellar region at about the same depth as c. CPU, caudate-putamen; CB, cerebellum; CC, cingulatecortex; CAI, hippocampal area; EC, entorhinal cortex; LS, lateral septum; P, Purkinje cells; R, raphe nuclei; ST, subthalamic nuclei.

cells of the cerebellum. Weaker signals were also detected inthe hippocampus, raphe nuclei, lateral septum, subthalamicnuclei, cingulate cortex, and entorhinal cortex. In a controlexperiment performed under the same conditions with plas-mid DNA fragments instead ofthe 5HT1B genomic fragment,no hybridization was observed.

DISCUSSIONOur binding and cyclase data indicate that the genomic clonethat we have isolated encodes a functional mouse 5HT1Breceptor. This receptor exhibits 59% amino acid identity withthe recently cloned human 5HT1D receptor (7). However, itis unlikely that "our 5HT1B receptor" is the mouse coun-terpart of that 5HT1D receptor, because, in the cases knownso far, receptor subtypes exhibit a higher degree of conser-vation across mammalian species-usually 85-95% aminoacid identity. Furthermore, we have recently cloned anothermouse receptor that is 89o homologous to the cloned 5HT1Dreceptor (results not shown) and that is therefore most likelythe mouse counterpart of the 5HT1D receptor. Our resultssuggest therefore that the 5HT1B/5HT1D receptors consti-tute a heterogenous family consisting of at least two differentreceptors in each mammalian species.The pattern of expression of the 5HT1B mRNA indicates

that a large proportion of the 5HT1B receptors are locatedpostsynaptically with respect to the afferent 5-HT-ergic fi-bers originating in the raphe nuclei. Only the transcriptsexpressed in the raphe nuclei (Fig. 4 a and b) corresponds to

receptors localized on the 5-HT-ergic neurons or autorecep-tors. Such autoreceptors have been proposed to modulate therelease of5-HT from the terminals of 5-HT-ergic neurons (fora review, see ref. 1).The main sites of expression of 5HT1B mRNA are the

striatum (caudate-putamen) and the Purkinje cells of thecerebellum (Fig. 4). The comparison between this mRNApattern and the pattern of the 5HT1B binding sites deter-mined by autoradiography is indicative ofthe pre- or postsyn-aptic localization of this receptor. The 5HT1B binding siteswere found mostly in the globus pallidus and substantia nigra(10) and to a lesser extent in the caudate-putamen. Incontrast, 5HT1B transcripts were detected in the caudate-putamen but not in the globus pallidus and substantia nigra.Since most efferent fibers from the caudate-putamen projectto the globus pallidus and substantia nigra, it is likely that the5HT1B receptors present in the globus pallidus and substan-tia nigra are localized presynaptically on the terminals ofprojecting striatal neurons. In good agreement with thishypothesis is the observation that experimental or patholog-ical lesions of the caudate-putamen, such as those observedin Huntington chorea (11, 12), result in a decrease in5HT1B/1D sites not only in the caudate-putamen but also inthe globus pallidus and substantia nigra. The presynapticlocalization of these 5HT1B receptors suggests that theymight modulate the release of neurotransmitters such asy-aminobutyric acid (GABA), substance P, or enkephalinsfrom striatal neuron terminals. Interestingly, the dopamineD1 receptors that have been reported to stimulate GABA

Neurobiology: Maroteaux et al.

3024 Neurobiology: Maroteaux et al.

release in the substantia nigra (13) are also localized presyn-aptically on projecting striatal neurons (14). Since D1 recep-tors activate adenylate cyclase, whereas 5HT1B receptorsare negatively coupled to adenylate cyclase, it is possible thatthese two receptors modulate GABA release in oppositeways in the substantia nigra.The other main site of expression of 5HT1B transcripts is

the Purkinje cells. 5HT1B binding sites have been detected inthe deep nuclei of the cerebellum that contain the Purkinjecell terminals (10). These 5HT1B sites could therefore cor-respond again to presynaptic receptors modulating GABArelease from Purkinje cells.

Expression of 5HT1B receptors in the striatum, subtha-lamic nuclei and cerebellum, which are brain structuresinvolved predominantly in movement control, suggests a rolefor these receptors in motor function. Some experimentaldata, such as an increase in locomotor activity after admin-istration of 5HT1B agonists to rats (15), support this hypoth-esis. Therefore, it might be interesting to investigate theeffects of 5HT1B selective drugs in the treatment of motordisorders of the striatum and cerebellum, such as Huntingtonchorea and cerebellar ataxias. The availability of a cell lineexpressing high levels of this receptor subtype should facil-itate the development of new 5HT1B agonists and antago-nists. The 5HT1B genomic clone will also allow us by meansof gene-targeting techniques, to produce mouse mutants andto analyze the consequences of these mutations on thephysiology of the animal.

Note Added In Proof. A rat 5HT1B receptor homologous to the mouse5HT1B receptor has been recently described (16).

L.M. and F.S. contributed equally to this work. We thank A. Stauband F. Ruffenach for making the oligonucleotides and D. Kauffmannand E. Rauscher for typing the manuscript. We are grateful to G.

Gombos, J. de Barry, and C. Mendelsohn for valuable discussions.This work was supported by grants from the Centre National de laRecherche Scientifique, the Institut National de la Sante et de laRecherche Medicale, the Association pour la Recherche contre leCancer, and Rh6ne-Poulenc Rorer.

1. Frazer, A., Maayani, S. & Wolfe, B. B. (1990) Annu. Rev.Pharmacol. Toxicol. 30, 307-348.

2. Julius, D. (1991) Annu. Rev. Neurosci. 14, 335-360.3. Green, S., Isseman, I. & Sheer, E. (1988) Nucleic Acids Res.

16, 369-370.4. Amlaiky, N. & Caron, M. G. (1985) J. Biol. Chem. 260,

1983-1986.5. Waeber, C., Schoeffter, P., Palacios, J. M. & Hoyer, D. (1989)

Naunyn-Schmiedeberg's Arch. Pharmacol. 340, 479-485.6. Hafen, E., Levine, M., Garber, R. L. & Gehring, W. J. (1983)

EMBO J. 2, 617-623.7. Hamblin, M. W. & Metcalf, M. A. (1991) Mol. Pharmacol. 40,

143-148.8. Fargin, A., Raymond, J. R., Lohse, M. J., Kolbika, B. K.,

Caron, M. G. & Lefkowitz, R. J. (1988) Nature (London) 335,358-360.

9. Schoeffier, P. & Hoyer, D. (1989) Naunyn-Schmiedeberg'sArch. Pharmacol. 340, 285-292.

10. Pazos, A. & Palacios, J. M. (1985) Brain Res. 346, 205-230.11. Waeber, C., Zhang, L. & Palacios, J. M. (1990) Brain Res. 528,

197-206.12. Waeber, C. & Palacios, J. M. (1989) Neuroscience 32, 337-347.13. Reubi, J. C., Iversen, L. & Jessel, T. (1977) Nature (London)

258, 652-654.14. Fremeau, R. T., Duncan, G. E., Fornaretto, M. G., Dearry,

A., Gingrich, J. A., Breese, G. R. & Caron, M. G. (1991) Proc.Natl. Acad. Sci. USA 88, 3772-3776.

15. Oberlander, C., Demassey, Y., Verdu, A., van de Velde, D. &Bardeley, C. (1987) Eur. J. Pharmacol. 139, 205-214.

16. Voight, M. M., Laurie, J. L., Seeburg, P. H. & Bach, A. (1991)EMBO J. 10, 4017-4023.

Proc. Natl. Acad. Sci. USA 89 (1992)