Journal of Neuroche,nist~’Lippincott—Raven Publishers, Philadelphia© 1997 International Society for Neurochemistry

Cloning and Expression of a Human Serotonin5-HT4 Receptor cDNA

use Van den Wyngaert, l~WalterGommeren, *peter Verhasselt, lMirek Jurzak,tJosée Leysen, Walter Luyten, and Eckhard Bender

Departments of Experimental Molecular Biology, *Applied Molecular Biology, and

tBiochemical Pharmacology, Janssen Research Foundation, Beerse, Belgium

Abstract: Using a combination of library screening andnested PCR based on a partial human serotonin 5-HT4receptor sequence, we have cloned the complete codingregion for a human 5-HT4 receptor. The sequence showsextensive similarity to the published porcine 5-HT4A andrat 5-HT4L receptor cDNA; however, in comparison withthe latter, we find an open reading frame correspondingto only 388 amino acids instead of 406 amino acids. Thisdifference is due to a frame shift caused by an additionalcytosine found in the human sequence after position1,154. Moreover, we also found the same additional cyto-sine in the rat 5-HT4 sequence. We confirmed the occur-rence of the sequence by examining this part of the se-quence in genomic DNA of 10 human volunteers and inrat genomic DNA. Based on a part of the genomic 5-HT4 receptor sequence that was identified in the cloningprocess, there seem to be at least two possible splicesites in the coding region of the gene. The human 5-HT4receptor, transiently expressed in COS-7 cells, showedradioligand binding properties similar to 5-HT4 receptorsin guinea pig striatal tissue. [

3H]GR113808 revealed KDvalues of 0.15 ±0.01 nMforthe human receptor and 0.3±0.1 nM in the guinea pig tissue. Binding constantsweredetermined for four investigated 5-HT

4 antagonists andthree agonists, and appropriate binding inhibition con-stants were found in each case. Stimulation of trans-fected COS-7 cells with 5-HT4-specific agonists causedan increase in cyclic AMP levels. Key Words: Seroto-nin—5-HT4 receptor—Cloning-—Human—Rat.J. Neurochem. 69, 181 0—1819 (1997).

Serotonin 5-HT4 receptors were first defined in pri-mary cultures of mouse embryonic colliculi, based onobserved 5-HT-induced cyclic AMP formation (Du-muis et al., 1988a,b). A series of benzamides knownfor their gastrokinetic properties mimicked the 5-HT-induced cyclic AMP formation in the colliculi neurons,whereby the potencies correlated with effects on intes-tinal tissue (Dumuis et al., 1989). Hence, it was sug-gested that the therapeutically interesting gastrokineticproperties of drugs like cisapride (Schuurkes et al.,1985) could be mediated by agonistic action on gut 5-

HT4 receptors (Briejer et al., 1995). The advent ofselective 5-HT4 antagonists (Gaster and Sanger, 1994;Sanger and Gaster, 1994) and of a radiolabeled 5-HT4antagonist [

3H]GR 113808 (Ford and Clarke, 1993)facilitated the further exploration of the 5-HT

4 recep-tors in the brain and in peripheral tissues (Ford andClarke, 1993; Eglen et al., 1995). The occurrence of5-HT4 receptors, demonstrated in mammalian brain(including human) using radioligand binding in re-gions of the extrapyramidal motor system of the basalganglia and the mesolimbic system, has suggested apossible role of 5-HT4 receptors in affectiveand cogni-tive processes (Waeber et al., 1993; Eglen et al.,1995). By using functional studies in isolated periph-eral tissues, 5-HT4 receptor-mediated responses weredetected in the gastrointestinal tract (Ford and Clarke,1993), the atrium (Kaumann et al., 1996), the bladder(Candura et al., 1996), and the pulmonary vein (Cocksand Arnold, 1992). Ullmer et al. (1995) and Schmucket al. (1996) showed expression of the rat, porcine,and human 5-HT4 receptor in blood vessels withoutincluding functional studies. The best characterizedproperty of 5-HT4 ligands probably concerns gastroki-netic effects. However, studies in various species sug-gest that there may be species differences with respectto agonistic and antagonistic properties of ligands. Toallow appropriate investigation of the properties of li-gands for therapeutic applications, cloning and expres-sion of human 5-HT4 receptors are of prime impor-tance.

Presently, six complete or partial cDNA sequences

Received April 24, 1997; revised manuscript received June 30,1997; accepted June 30, 1997.

Address correspondence and reprint requests to Dr. E. Bender atDepartment of Experimental Molecular Biology, Janssen ResearchFoundation, Tumhoutseweg 30, B-2340 Beerse, Belgium.

Abbreviations used: AC, adenylyl cyclase; DMEM, Duibecco’smodified Eagle medium; DMSO, dimethyl sulfoxide; 5-HT, 5-hy-droxytryptamine, serotonin; K,, inhibition constant; 5-MeOT, 5-methoxytryptamine; ON1 —14, oligonucleotides 1—14; ORF, openreading frame; plC50, negative logarithm of the concentration thatinhibits 50% of specific binding by the radioligand.

1810

CLONING OF A HUMAN 5-HT4 RECEPTOR cDNA 1811

encoding 5-HT4 receptors are known: two splice vari-ants from the rat (Gerald et al., 1995), one murinesequence (Claeysen et al., 1996), two partial porcinesplice variants (Ullmer et al., 1995), and a partialhuman sequence (Ullmer et al., 1995). The predictedprotein structures reveal seven transmembrane do-mains for the complete open reading frames (ORFs).According to their structure and 5-HT4 receptor-cou-pled signal transduction events (increase in cyclicAMP formation, opening of K + channels), 5-HT4 re-ceptors have been classified as G protein-coupled re-ceptors. To obtain a 5-HT4 receptor eDNA from ahuman brain eDNA library, we used a combination oflibrary screening and nested PCR. The derived humansequence was highly similar to the published rat 5-HT4L sequence (Gerald et al., 1995); however, due toa frame-shift in the 3’ region, a different stop codonis used to terminate translation. Therefore, the humanand rat sequences were confirmed by sequencing corre-sponding PCR samples from genomic DNA fromblood of normal human volunteers and rats. In bothspecies, the found frame shift was confirmed. The hu-man cDNA was subcloned into the expression vectorpcDNA3 and transfected transiently into COS-7 cellsfor the investigation of the pharmacological profile andsecond messenger coupling of the receptor. Radioli-gand binding properties of the cloned human receptorwere compared with those of 5-HT4 receptors in guineapig striatal tissue. The pharmacological profile of thereceptor was evaluated further in light of the functionalproperties of ligands.

MATERIALS AND METHODS

MaterialsdNTPs, MgC12, and PCR buffer II were obtained from

Perkin—Elmer Cetus (Foster City, CA, U.S.A.). Pfu poly-merase and the Robocycler PCR block were products ofStratagene (La Jolla, CA, U.S.A.). TaKaRaEx Taq polymer-ase was obtained from TaKaRa (Shigu, Japan). TaqStartantibody and human brain 5 ‘-STRETCH PLUS eDNA li-brary (XgtlO) were from Clontech (Palo Alto, CA, U.S.A.).T4 DNA ligase and restriction endonucleases were productsof Boehringer (Mannheim, Germany). The MultiprimeDNA labeling system, Hybond-N filters, and [

3H]GR113808 with a specific activity of 3.07 Tbq/mmol were ob-tained from Amersham (Little Chalfont, U.K.). [32PIdCTPwas purchased from NEN Du Pont (Wilmington, DE,U.S.A.). The Qiagen Lambda Maxi kit, plasmid preparationkits, andthe QiaquickPCR purification kit were from Qiagen(Hilden, Germany). The PRISM Ready Reaction Dye Ter-minator Cycle Sequencing Kit and the ABI 377 or 373Asequencing machines were from AppliedBiosystems (FosterCity, CA, U.S.A.). The GeneAmp PCR System 9600 wasfrom Perkin—Elmer (Norwalk, CT, U.S.A.). The MicroSpinG-50 columns were obtained from Pharmacia Biotech(Upp-sala, Sweden), and the mammalian expression vectorpcDNA3 was obtained from Invitrogen (Carlsbad, CA,U.S.A.). Dulbecco’s modified Eagle medium (DMEM) andfetal calf serum were from Life Technologies (Gaithersburg,MD, U.S.A.). Wistar rats were from Charles River Labora-

tories (Sulzfeld, Germany). The Bradford protein assay wasperformed with the reagent supplied from Bio-Rad (Naza-reth Eke, Belgium). The NEN flash plate assay was suppliedby Du Pont de Nemours (Brussels, Belgium). The liquidscintillation spectrometer and the scintillation fluid UltimaGold MV were from Packard (Meriden, CT, U.S.A.). Cisa-pride was synthesized by Janssen Pharmaceutica NV(Beerse, Belgium). Tropisetron wasa gift from Sandoz (Ba-sel, Switzerland), and GR 125487 was a gift from Glaxo(Ware, U.K.). SB 204070 and GR 113808 were synthesizedby Janssen Pharmaceutica for its own purposes. All com-pounds were dissolved and diluted in dimethyl sulfoxide(DMSO; except the indoleamines, which were dissolved inwater and protected from light throughout the experiment).The final DMSO concentration in the tests did not exceed0.5% (vol/vol). The GraphPad Prism program was fromGraphPad Software, Inc. (San Diego, CA, U.S.A.).

General molecular biological methodsUnless otherwise indicated, all PCR reactions were per-

formed in a total volume of 50 ~.tl,containing 1 p~lof eDNAand 2.5 U of Pfu polymerase in 1 x PCR buffer II, 200 ,uMdNTPs, 200 nM primers, and 2.0 mM MgCl

2. PCR condi-tions were 5 mm of denaturation at 95°C, 1 mm at 54°C,and 2 mm at 75°C, followed by a 10-mm incubation at75°C.DNA manipulations were done according to standardprotocols (Maniatis et al., 1982).

DNA sequencing was carried out with reagents from thePRISM Ready Reaction Dye Terminator Kit and run on aGeneAmp PCR System 9600 according to the specificationsof the supplier.

Preparation of human 5-HT4 receptor probe andeDNA library screening

Two primers [oligonucleotide 1 (ON1): 5’-CTG CTGCCA GCC TTT GGT CTA T-3’; 0N2: 5 ‘-TCT CTG TCCTCA TGC GAT GAG TG-3 ‘] were designed based on thepublished partial human 5-HT4 receptor eDNA sequence(Ullmer et al., 1995). PCR reactions were performed usingthe human brain XgtlO 5 ‘-STRETCH PLUS eDNA library.After the 398-bp PCR product was subcloned into pUCl8,a 370-bp human 5-HT4 eDNA insert was removed from thepUC18 construct by restriction digest using StuI and KpnI.TheeDNA wasthen randomly labeled with [

32P]dCTP usingthe Multiprime DNA labeling system. A human brain kgtl05 ‘-STRETCH PLUS eDNA library was screened as de-scribed in the Lambda Library Protocol Handbook (Clon-tech). One positive plaque was observed, and after tworounds of plaque purification a single positive plaque (clone17) was cored out. Sequencing by primer walking of the Xclone 17 insert revealed that it contained part of the human5-HT

4receptor eDNA, from nucleotide position 507 to posi-tion 1,077 of the ORF.

Nested PCR amplification and cloning of the 5’and 3’ ends of the human 5-HT4 ORE

Based on the sequence of the X clone 17 insert, four oligo-nucleotides were designed (0N5: 5 ‘-AAG GCC ACC ACAGAG CAG-3’; 0N6: 5 ‘-TGC CCG TTG TAA CAT CTGAA-3’; 0N7: 5’-TGT TCA ACC ACA ACC ATT AATGG-3’; 0N8: 5’-ATA GAC CAA AGG CTG GCA GC-3’). PCR reactions combining these oligonucleotides withone universal )~gtl0primer (0N3: 5 ‘-AGC AAG TTC AGCCTG GTT AAG T-3’; 0N4: 5 ‘-TTA TGA GTA TTT CTTCCA GGG-3’) were performed on the ~gtl0 human brain

J. Neurochem., Vol. 69, No. 5, 1997

1812 I. VAN DEN WYNGAERT ET AL.

5 ‘-STRETCH PLUS eDNA library to amplify the missing5’ and 3’ parts of the human 5-HT4 ORF by nested PCR.To obtain the 5’ region, primers 0N3 and 0N6 were usedin a first round of PCR, followed by a secondary PCR using1 p~lof the resulting product in combination with primers0N3 and ONS. In an analogous fashion, the missing 3’region was amplified by combining first 0N4 and 0N8 andusing the resulting product as substrate for the primer combi-nation 0N4 and ON7. The PCR was performed under stan-dard conditions, except that TaKaRa Ex Taq polymerase,(1.25 U) was used. The reactions were cycled 35 times for30 s at 95°C,30 s at 54, 55, 56, 57, 58, or 59°C,and 2 mmat 72°Cin the Robocycler PCR block, followed by a 1 0-mmincubation at 72°C.PCR products were ligated into pUC18.

Assembly of the full-length human 5-HT4receptor coding region

The full-length eDNA was assembled by a two-step liga-tion. In the first step, the EcoRI/PfIMI restriction enzymefragment of nestedPCR product 0N3/0N5 wasligated withthe XbaI/PflMI restriction enzymefragment ONl /0N2 intothe EcoRIIXbaI large fragment of the mammalian expres-sion vector pcDNA3, yielding pcDNA3/ON1ION2/0N310N5. In parallel, the EcoRIIAf1111 restriction enzyme frag-ment of the X clone 17 insert was ligated together with theBstUIIAf1111 restriction enzyme fragment of the nested PCRproduct 0N4/0N7 into the EcoRIIEcoRV large fragmentofpcDNA3, yielding pcDNA3I l7/0N4/0N7. In thesecondstep, the EcoRIISphI small fragment of pcDNA3 /ON 1 /0N2/0N3IONS was ligated together with the SphI/NotIsmall fragment of pcDNA3 /1 7/0N4/0N7 into the EcoRIINotI large fragment of pcDNA3.

Sequence verification of the human and rat 5-HT4ORF 3’ end

To verify the nucleotide sequence at the 3’ end of thehuman and rat 5-HT4 coding region, control PCRs were setup on genomic DNA to amplify the 3’ end as follows: twooligonucleotides (0N9: 5’-AGA GTC AGT GTC ACCCGC CAG-3’; ON1O: 5 ‘-AAG CAG CAGCTT AGG ACCTG-3’) were designed downstreamof theputative splice site(assigned on the basis of the reported rat 5-HT4 receptorsequence). PCR was carried out on genomic DNA samplesprepared from the blood of 10 human volunteers in a volumeof 100 ftl using standard conditions.

Similarly, PCR was carried out on genomic DNA preparedfrom rat blood to amplify the short and long rat 5-HT4 splicevariants. This was performed using oligonucleotides ONll(5’-TTT GCA TAG TGG TCA ACA CCA GG-3’), 0N12(5’-CGT CTT CAA TCA AAA GCA TGA TTC C-3’),0N13 (5’-AGA GTC GGT GTC ACC TCA CAG-3’), and0N14 (5’-CAG CTT AGG ACT GGC TTC TTT TC-3’).

Expression of the human 5-HT4 receptor inmammalian cells

COS-7 cells were grown in DMEM supplemented with10% fetal calf serum. A large-scale plasmid preparation of5-HT4/pcDNA3 was made using the Qiagen large-scaleplasmid prep kit. Plasmid DNA was transfected into COS-7 cells (seeded at 30,000 cells/cm

2 the day before transfec-tion) using the DEAE-dextran method (Huylebroeck et al.,1988) in serum-free DMEM for 30 mm. The transfectionmedium was removed and replaced by COS-7 medium sup-plemented with 100 ~iM chloroquine. After 4 h, the chlo-roquine medium was replaced by COS-7 medium containing

4 mM sodium butyrate. Sixteen hours later, this mediumwas replaced by COS-medium containing only 2% fetal calfserum. Incubation was continued for another 48 h beforethe cells were harvested for binding and signal transductionassays.

Membrane preparationThe transfected COS-7 cells were cultured on 150-mm

Petri dishes and washed twice with ice-cold phosphate-buf-fered saline. The cells were then scraped from the plateswith a cell scraper, suspended in 50 mM Tris-HC1 buffer,pH 7.4, andharvested by centrifugation for 10 mm at 16,000g. The pellet was resuspended in 5 mM Tris-HC1, pH 7.4,and homogenized with an UltraTurrax homogenizer; the re-sulting membranes were collected by centrifugation for 20mm at 25,000 g. Membranes were stored at —70°C in 50mM Tris-HC1 buffer, pH 7.4, at a protein concentration of1 mg/ml. For the preparation of membranes from guinea pigstriatum, the tissue was pooled in cold 50 mM Tris-HC1buffer, pH 7.4, homogenized, centrifuged, and collected asdescribed for the cell membranes. The Bradford proteinassay was used for protein determination with bovine serumalbumin as a standard.

Radioligand bindingAssay mixtures (0.5 ml) contained [3H]GR 113808, an

aliquot of membrane preparation (25—75 ,ug of protein fortransfected COS-7 cells, 200—300 ,ug of protein for guineapig striatal tissue), and solvent for total binding or 10 pM5-HT to determine nonspecific binding. The assay bufferconsisted of 50 mM HEPES/NaOH, pH 7.5, 1 ,aM pargyline(monoamine oxidase inhibitor), and paroxetine (5-HT trans-porter inhibitor), and the mixture was incubated at 37°Cfor30 mm. Competition binding experiments were performedwith a 0.2 nM concentration of the radioligand. Ligand con-centration binding isotherms were obtained using 10 concen-trations of [3H]GR 113808 in a concentration range of 16pM to 1.6 nM. The incubation was terminated by rapidfiltration over Whatman GF/B filters presoaked in 0.1%polyethylenimine, and three washing steps with 3 ml of ice-cold Tris-HC1 buffer, pH 7.4. Filter-bound radioactivity wascounted in a liquid scintillation spectrometer using 2 ml ofscintillation fluid.

Ligand concentration binding isotherms (rectangular hy-perbola) and sigmoidal inhibition curves were calculated bynonlinear regression analysis according to algorithms de-scribed by Oestreicher and Pinto (1987). The maximal num-ber of binding sites (Bmax) and equilibrium dissociation con-stant (KD) of the radioligand and the plC

50 (negative loga-rithm of the concentration that inhibits 50% of specificbinding by the radioligand) values of competitors were de-rived from the curve fitting. Apparent inhibition constant(K~)values were calculated according to the equation ofCheng and Prusoff (1973). Graphs were prepared using theGraphPad Prism program.

Measurement of cyclic AMP formationFor the stimulation of adenylyl cyclase (AC), COS-7 cells

were transfected andcultured in 24-multiwell plates (seedingdensity of 30,000/well). The DMEMcontained 1% dialyzedcalf serum (<0.1 nM 5-HT). On the third day followingtransfection, the cells were washed with buffer consisting of2S mM Tris-HC1, pH 7.4, 120 mM NaCl, 5 mM KCI, 0.8mM MgC12, 1.8 mM CaC12, 15 mM glucose, and 0.04 mMphenol red. Isobutylmethylxanthine at 1 mM, I 1iM pargy-

J. Neurochem., Vol. 69, No. 5, 1997

CLONING OF A HUMAN 5-HT4 RECEPTOR cDNA 1813

line, and 1 ~.sMparoxetine were then added to the buffer. Toestimate themaximal AC stimulation, 100 jiM forskolin wasused. The incubation with compounds was performed for 20mm at 37°C.The final concentration of DMSO (wheneverneeded to dissolve the compounds) did not exceed 0.5%(vol/vol) and was also included in thecorresponding controlsamples. All points were from triplicate assays. The incuba-tion was stopped by acidification with 0.1 ml of ice-cold 1M HC1O4, followed by neutralization with K2HPO4. Afterthe formation of a KC1O4 precipitate, the plates were centri-fuged for S mm at 2,000 g and the supernatant was assayedfor cyclic AMP content using the NEN flash plate assayaccording to the protocol of the supplier.

RESULTS

Based on the partial eDNA sequence of the human5-HT4 receptor (Ullmer et al., 1995), two PCR primerswere designed to amplify this 397-bp-long fragmentof human 5-HT4 eDNA, stretching from position 366to position 762 of the ORF [following the alignmentof the published rat eDNA sequence (Gerald et al.,1995)1. This PCR product was cloned, radiolabeled,and then used to screen a commercial human braineDNA library, resulting in one positive clone (clone17). Sequence analysis of the 1,650-bp-long insertshowed extensive similarity to the published rat andpig sequences from position 507 to position 1,077 ofthe rat 5-HT4 ORF (Fig. 1). However, downstreamof position 1,077 and upstream of position 507, nosignificant similarity was detectable (Fig. 1). As theknown rat (Gerald et al., 1995) and porcine (Ullmeret al., 1995) splice variants diverge at these points andthe sequences at these positions are in good agreementwith the consensus 5’ and 3’ splice sites (Fig. 1; Sha-piro and Senapathy, 1987), the insert of clone 17 ismost likely a genomie DNA fragment, contaminatingthe eDNA library.

To obtain the remaining parts of the human 5-HT4eDNA sequence, primers were designed based on the5’ and 3’ ends of the sequence of clone 17 that stillshowed sequence similarity to the rat 5-HT4 eDNA.These primers were used in combination with primerscomplementary to the XgtlO vector sequence 5’ and3’to the cloning site in a nested PCR reaction.

The sequence of the 0N3/0N5 PCR product cone-

FIG. 1. Nucleotide sequence comparison ofthe insertfrom clone 17 with the porcine 5-HT4A (UlImer et al.,1995) and rat 5-HT4L (Gerald et al., 1995) sequences.The positions where the sequences diverge are indi-cated (A). Sequences that correspond to the 3’ or5’ splice consensus site are boxed. Numbering of thenucleotides is according to their position in the ORF;for the porcine partial sequence, it is based on thehuman 5-HT4 ORE.

sponds to the 5’ region of the human 5-HT4 receptoreDNA from position —289 to position +386, with theA of the start eodon being position + 1. The 0N4/0N7PCR product corresponds to the 3’ region of the human5-HT4 receptor eDNA from position + 1,057 up to thestop eodon and continuing for 301 nueleotides into the3’ untranslated region. Therefore, the resulting prod-ucts were shown to contain the missing parts of thehuman 5-HT4 coding region, as well as part of the 5’and 3’ untranslated regions. The full-length nueleotidesequence of the coding region and derived amino acidsequence are shown in Fig. 2.

The human sequence showed overall similarity tothe published rat 5-HT4 receptor long splice variantsequence (Fig. 3), but it differed by one additionalcytosine after position 1,154 of the rat coding sequence(Fig. 2). This reading frame shift resulted in a com-pletely different amino acid sequence downstream andusage of a different stop eodon, terminating the proteinafter amino acid position 388 instead of amino acid406. To verify the human sequence, the 3’ region of theORF was amplified from genomic DNA of 10 differenthuman volunteers. All samples showed the additionalcytosine after position 1,154.

To determine whether this was a species difference,the long splice variant 3’ part of the rat 5-HT4 receptorwas amplified from two different rat genomie DNAsamples. In contrast to the published sequence (Geraldet al., 1995), we also found a cytosine triplet insteadof a doublet present in the rat sequence. When theadditional eytosine is inserted into the rat nueleotidesequence, the derived rat long splice variant is onlyone amino acid longer than the short splice variant.This shortened rat 5-HT4L splice variant displays anORF that corresponds closely to the cloned human 5-HT4 receptor (see Fig. 3).

The pharmacological characteristics of the human 5-HT4 receptor protein were investigated by radioligandbinding assays on membranes from COS-7 cells tran-siently transfeeted with the human 5-HT4/peDNA3 ex-pression plasmid. Preliminary functional data were ob-tained by measuring 5-HT-induced cyclic AMP forma-tion in the same cell population.

Ligand concentration binding isotherms with [3H]-

J. Neurochem., Vol. 69, No. 5, 1997

1814 I. VAN DEN WYNGAERT ET AL.

GR 113808 on membranes from three independenttransfeetions in COS-7 cells revealedBmax values rang-ing from 1,114 to 3,225 fmol/mg of protein and amean K

0 value of 0.15 ±0.1 nM. The correspondingvalues on preparations from the guinea pig striatumwere 182 ±30 fmol/mg of protein and 0.3 ±0.1 nMfor the Bmax and K0 values, respectively (n = 3). Astraight line in a Seatehard analysis suggested thepres-ence of a single high-affinity binding site in both tis-sues. Representative curves are shown in Fig. 4. Mem-branes from COS-7 cells transfected with the pcDNA3vector control did not show any specific [

3HIGR113808 binding (i.e., inhibited by 5-HT).

Several selective ligands for the 5-HT4 receptor

were used in competition binding experiments as refer-ence compounds to define the pharmacological profileof the cloned human 5-HT4 receptor (Briejer et al.,1995). High-affinity antagonists with high selectivityfor the 5-HT4 receptors were used: SB 204070 (Wardleet al., 1994), GR 125487 (Gale et al., 1994), and GR113808 (Grossman et al., 1993), as well as the first-described, but weaker and nonseleetive 5-HT4 antago-nist tropisetron [ICS 205-930 (Dumuis et al., l988b)].In addition, we investigated 5-HT, its natural derivative5-methoxytryptamine (5-MeOT), and the benzamidegastroprokinetie agent cisapride, known to be a partialagonist of the receptor. Inhibition curves of radioligandbinding in membranes from transfeeted COS-7 cellsand from guinea pig striatum are depicted in Fig. 5,and the mean plC50 and K values are presented inTable 1.

The functional coupling of the cloned human recep-tor in COS-7 cells to the cyclic AMP pathway wasinvestigated in preliminary experiments. These experi-ments have clearly shown a stimulation of the ACsystem by receptor-specific agonists (Fig. 6). Basalcyclic AMP levels were 3—4 nM in the presence of<0.1 nM 5-HT. Incubation with 5-HT, 5-MeOT, orcisapride increased the levels of cyclic AMP up tosevenfold. In veetor-transfected cells, cyclic AMP lev-els were not elevated following incubation with the 5-HT4 agonists (data not shown). The forskolin-inducedcyclic AMP level in 5-HT4 receptor transfeeted cellsremained unchanged after 5-HT addition. Hence, wefound no indication for an inhibitory action of 5-HTon the cyclic AMP system.

DISCUSSION

On the basis of the published partial human 5-HT4eDNA sequence, we attempted to clone a full-length

FIG. 2. Nucleotide and deduced amino acid sequence of thecloned human 5-HT4 receptor. Potential N-linked glycosylation(•), protein kinase C phosphorylation (V), and palmitoylation(~)sites are indicated. Positions of amino acids and nucleotidesare given at the end ofeach line. The additional C that was foundcompared with the published rat 5-HT4L cDNA (Gerald et al.,1995) is underlined.

J. Neurochem., Vol. 69, No. 5, 1997

CLONING OF A HUMAN 5-HT4 RECEPTOR cDNA 1815

FIG. 3. Alignment of the six known partial [pig5-HT4A, pig 5-HI45 (UlImer et al., 1995), and rat5-HT4L (this study)) and complete [rat5-HT45 (Ger-ald et al., 1995), mouse 5-HT4L (Claeysen et al.,1996), rat 5-HT4L (Gerald et al., 1995), and a human5-HT4 (this study)] 5-HT4 receptor amino acid se-quences. The putative transmembrane domains(TM) are boxed.

eDNA by screening a commercially available humanbrain eDNA library. However, the only clone that wasfound contained a fragment of genomie DNA as aninsert. To some extent, this parallels the difficultiesdescribed in finding a full-length rat eDNA clone infour different rat eDNA libraries (Gerald et al., 1995),

most likely reflecting the low abundance of the 5-HT4receptor mRNA.

The central sequence of the insert from our genomieclone shows a high degree of sequence similarity tothe published rat 5-HT4 eDNA; however, at the 5’ and3’ ends of this central segment, the sequence similarity

.1. Neurochem., Vol. 69, No. 5, 1997

1816 I. VAN DEN WYNGAERT ET AL.

FIG. 4. [3HJGR113808 concentration binding isotherm formembrane preparations from COS-7 cells transfected with thehuman 5-HT

4 receptor (A) and from guinea pig striatum (B).Total binding (•), nonspecific binding (•), and specific binding(D) are presented as means of duplicate determinations. Thedata represent a typical experiment out of three independentexperiments. Nonspecific binding was determined in the pres-ence of 10 jiM 5-HT. KD and ~ values were derived from curvefitting using nonlinear regression analysis. The mean values forthe 5-HT4 transfected COS-7 cells were a KD of 0.15 ±0.01 nMand a ~ of 2,221 fmol/mg of membrane protein; for the guineapig striatum, a K0 of 0.3 ±0.1 nM and a Brna~of 182 ±30 fmol/mg of membrane protein were determined.

stops (Fig. 1). The 5’ site of sequence divergencecorresponds to the position in the porcine 5-HT4Bwhere an insertion is found, suggesting that an intron/exon boundary is present at this point in the humangene as well. At the 3’ end, the similarity stops wherethe rat long and short splice variants diverge in se-quence; therefore, this point also corresponds to anexon/intron boundary. This interpretation is strength-ened by the occurrence of sequences closely resem-bling splice consensus motifs at the site of sequencedivergence (Fig. 1; Shapiro and Senapathy, 1987).

The missing parts of the ORF at the 5’ and 3’ endswere cloned by nested PCR performed on a humanbrain eDNA library. The nucleic acid and derivedamino acid sequences showed a high degree of similar-ity to the published rat 5-HT4 long splice variant; how-ever, we also observed an additional cytosine afterposition 1,154. The recently cloned murine 5-HT4 re-ceptor eDNA (Claeysen et al., 1996) is to a largeextent similar to the published long splice variant ofthe rat (Fig. 3), showing only 16 differences at the

amino acid level and confirming also the eytosmne dou-blet after position 1,152.

To confirm that our sequence is correct at the 3’end, we examined this part of the sequence in genomicDNA from 10 different human volunteers. In all eases,we obtained our original sequence, suggesting that thisis the correct sequence. To investigate whether thisfinding pointed toward a species difference, we ampli-fied the same region from rat genomie DNA. In twoindependent experiments, we found a sequence withan additional eytosine after nueleotide position 1,154,corresponding to our findings on the human receptor.The amino acid sequence of the long splice variant ofthe rat receptor derived from the nucleotide sequenceincluding the cytosine triplet is highly similar to ourhuman sequence (Fig. 2). However, this indicates thatour long rat splice variant is only 388 amino acids longinstead of 406. Further studies are needed to determinewhether the observed difference is based on a variantDNA sequence or due to a sequencing error. Two par-tial 5-HT4 receptor eDNA splice variants are knownin the pig, where the alternative splicing occurs at adifferent position from that in the rat sequence. Regard-ing this variation, our clone is similar to the 5-HT4Asequence (Fig. 3).

Taken together, the few known 5-HT4 receptoreDNA sequences from different species show a re-

FIG. 5. inhibition by 5-HT4 antagonists and agonists of [3H]GR

113808 (0.2 nM) binding to membrane preparations from COS-7 cells transfected with the human 5-HT

4 receptor (A) and fromguinea pig striatum (B). Depicted points are means of two tofive independent experiments, and the calculated plC~~valuesare given in Table 1.

J. Neurochem., Vol. 69, No. 5, 1997

CLONING OF A HUMAN 5-HT4 RECEPTOR cDNA 1817

TABLE I. Potency of compounds to compete with 0.2 nM [3H]GR 113808 binding in membranes from COS-7 cells

transiently transfected with the human 5-HT4 receptor and from guinea pig ,ctriatum

Compound

Human 5-HT4 receptor inCOS-7 cells

Guinea pig 5-HT4 receptor inthe striatum

Rat 5-HT4Lreceptor in

COS-7 cells”pK,

Human brainmembranes

0pK,plC

50 ±SD (n) pK1 plC50 ±SD (n) pK

AntagonistsSB 204070 9.7 ±0.2 (4) 10 10 ±0.3 (4) 10.1 ND NDGR 125487 9.5 ±0.1 (3) 9.8 9.6 ±0.2 (3) 9.7 ND NDGR 113808 9.4 ±0.03 (3) 9.6 9.3 ±0.3 (2) 9.4 ND 9.8Tropisetron 6.4 ±0.1 (3) 6.7 6.6 ±0.3 (2) 6.7 6.7 6.7

Agonists

Cisapride 6.9 ±0.2 (4) 7.2 7.1 ±0.1 (3) 7.2 6.9 6.85-HT 6.6 ±0.3 (5) 6.9 7.2 ±0.2 (3) 7.4 6.8 6.55-MeOT 5.7 ±0.2 (3) 6 6.5 ±0.2 (2) 6.6 6.4 6.3

Literature data obtained for the rat 5-HT4L receptor or on homogenates from human brain are included for comparison. ND, not determined.From Gerald et al., 1995;

0from Waeber et al., 1993.

markable sequence variation and various possibilitiesto generate splice variants. Detailed analysis of thespecies and tissue-dependent presence of these differ-ent variants will be important for the elucidation ofthe 5-HT

4 receptor physiology and pharmacology. Forthe rat long and short 5-HT4 receptor splice variant,different poteneies for AC stimulation have been de-scribed previously (Gerald et al., 1995). As we foundthe long splice variant to be actually only 388 aminoacids long, however, this cannot be attributed to afourth, additional phosphorylation site. Further supportfor the existence of 5-HT4 receptor variant sequencescan be drawn from reports on tissue-dependent desen-sitization rates of 5-HT4 receptor preparations (Ansa-nay et a]., 1992), which are attributed to differencesof the intercellular C-termini of the involved 5-HT4receptors (Gerald et al., 1995). The coupling effi-ciency of 5-HT4 receptors was also reported to be tis-sue-specific (Eglen et al., 1995), which is consistentwith a different GTP sensitivity of radioligand bindingin even closely related membrane preparations (Gross-man et al., 1993).

FIG. 6. Stimulation of AC in COS-7 cells transfected with thehuman 5-HT4 receptor by specific agonists. Basal levels andlevels after stimulation with a 1 jiM concentration of the indicatedcompounds from a triplicate determination are presented. cAMP,cyclic AMP.

In addition to the discussed interspecies sequencesimilarity, our pharmacological data strongly supportthe classification of the cloned human receptor into the5-HT4 receptor family. The high affinity of [

3H]GR113808 is a first line of evidence (Grossman et al.,1993). This radioligand has already proven its use-fulness in the mapping of 5-HT

4 receptors in the rat,guinea pig, and human brain (Grossman et al., 1993;Waeber et al., 1993). The K0 values measured in thisstudy on the human receptor expressed in COS-7 cellsand on the guinea pig striatal receptor (0.15 and 0.3nM, respectively) are in excellent accordance with areported K0 of 0.2 nM for the guinea pig striatumand human brain membranes (Grossman et al., 1993;Waeber et al., 1993). GR 113808 has also been re-ported to be a potent 5-HT4 receptor antagonist in theguinea pig ascending colon and rat esophagus withpA2 values of 9.3 and 9.5, respectively, and to have a>3,000-fold selectivity over other receptors (Gross-man et a]., 1993; Sanger and Gaster, 1994). The pK,value of cold GR 113808 in our binding tests on bothpreparations (9.4 and 9.6) is in good agreement withthese functional data.

The rank order of potency of the compounds usedfor the competition of [

3H]GR 113808 binding onCOS-7 cell membranes transfeeted with the 5-HT

4 re-ceptor clone (and on the guinea pig tissue as well)matches the expected pharmacological profile. The li-gand found to have the highest pK, values on thehumanand guinea pig receptor was SB 204070 (10 and 10.1,respectively). This benzodioxane derivative is themost potent and selective 5-HT4 antagonist describedso far (Gaster and Sanger, 1994) with an apparent pA2of 10.8 on the guinea pig distal colon. It has a 5,000-fold selectivity over all other receptor preparationstested (Wardle et al., 1994).

GR 125487 is a metabolically stable derivative ofGR 113808, with a prolonged in vivo biological half-life and slightly higher affinity (pK, = 10.4) on guineapig striatum (Gale et a]., 1994).

J. Neurochen,., Vol. 69. No. 5, 1997

1818 I. VAN DEN WYNGAERT ET AL

Another competitor tested, the potent 5-HT3 recep-tor antagonist tropisetron (ICS 205-930) was the firstdescribed 5-HT4 antagonist (Dumuis eta]., 1988b). Itsmoderate potency in binding tests on both membranepreparations (pK1 of 6.7 for the human and guinea pigreeepior) is in good agreement with the first reportsabout the inhibition of 5-HT-stimulated AC on mouseeollieular neurons [pK, = 6.15 (Ansanay et al., 1992)]and binding data (pK1 of 6.7) on human brain mem-branes (Waeber et a]., 1993). The tested agonists, in-cluding the full agonists 5-HT and 5-MeOT and thepartial agonist eisapride, exhibited lower affinities onboth membrane preparations in competition with a ra-diolabeled antagonist than reported in functional testson the guinea pig colon, where the pEC50 values were8.0, 7.8, and 7.5, respectively (Schuurkes et al., 1985;Leung et al., 1996). Agonists bind with high affinityto the G protein-coupled state of the receptor and havelower affinity for the uncoupled receptor. As radiola-beled antagonists detect both states of the receptor,agonist competition often shows an overall lower plC5ovalue, reflecting the affinity for the uncoupled receptor(Adham eta]., 1996). Consistent with this assumptionand our data, pK~values of 7.5 and 6.8 for 5-HT and5-MeOT, respectively, were observed in the guinea pigstriatum (Leung et al., 1996).

In summary, the pharmacological profile of thecloned human receptor matches the profiles .found inthe guinea pig striatum and human brain membranes(see Table 1; Waeber et a]., 1993), as well as that ofthe cloned rat 5-HT4L receptor expressed in COS-7cells (see Table 1; Gerald et al., 1995). The stimula-tory effect of 5-HT on cyclic AMP formation observedin 5-HT4/peDNA3 transfeeted COS-7 cells is an im-portant criterion of the definition of the 5-HT4 receptorfamily and is in agreement with many reports (Clarkeand Boekaert, 1993). The ability of the indole structure5-MeOT and the benzamide eisapride to stimulate cy-clic AMP production is a particular feature of the 5-HT4 receptor and has also been shown for the clonedrat 5-HT4L receptor (Gerald et al., 1995).

The characterization of a particular receptor is basedupon its structure, its second messenger coupling, andits pharmacological profile. The human receptor de-scribed in this report meets all the criteria to classifyit as a 5-HT4 receptor.

Acknowledgment: The authors thank Dr. Michel Briejerfor critical reading of the manuscript.

REFERENCES

Adham N., Gerald C., Schechter L., Vaysse P., Weinshank R., andBranchek T. (1996) [

3H]5-Hydroxytryptamine labels the ago-nist high affinity state of the cloned rat 5-HT

4 receptor. Eur. J.Pharmacol. 304, 2331—2335.

Ansanay H., Sebben M., Bockaert J., and Dumuis A. (1992) Charac-terization of homologous 5-hydroxytryptamine4 receptor desen-sitisation in colliculi neurons. Mol. Pharmacol. 42, 808—816.

Briejer M., Akkermans L. M. A., and Schuurkes J. A. J. (1995)Gas-

trointestinal prokinetic benzamides: the pharmacology underly-ing stimulation of motility. Pharmacol. Rev. 47, 631—651.

Candura S. M., Messori E., Franceschetti 0. P., D’Agostino G., Vi-cmi D., Tagliani M., and Tonini M. (1996) Neural 5-HT4 recep-tors in the human isolated detrusor muscle: effects of indole,benzimidazolone and substituted benzamide agonists and antag-onists. Br. J. Pharmacol. 118, 1965—1970.

Cheng Y.-C. and Prusoff W. H. (1973) Relationship between theinhibition constant (K,) and the concentration of inhibitor whichcauses 50 per cent inhibition (I~~)of an enzymatic reaction.Biochem. Pharmacol. 22, 3099—3108.

Claeysen S., Sebben M., Journot L., Bockaert J., and Dumuis A.(1996) Cloning, expression and pharmacology of the mouse 5-HT4L receptor. FEBS Lett. 398, 19—25.

Clarke D. E. and Bockaert J. (1993) 5-HT4 receptor: current status,in Serotonin (Vanhoutte P. M., Saxena P. R., Paoletti R., Bru-nello N., and Jackson A. S., eds), pp. 107—117. Kluwer Aca-demic Publishers, Cambridge, Massachusetts.

Cocks T. M. and Arnold P. J. (1992) 5-Hydroxytryptamine (5-HT)mediates potent relaxation in the sheep isolated pulmonary veinvia activation of 5-HT4 receptors. Br. J. Pharmacol. 107, 591—596.

Dumuis A., Bouhelal R., Sebben M., Cary R., and Bockaert J.(l988a) A nonclassical 5-hydroxytryptamine receptor posi-tively coupled with adenylate cyclase in the central nervoussystem. Mol. Pharmacol. 34, 880—887.

Dumuis A., Bouhelal R., Sebben M., and Bockaert J. (1988b) A 5-HT receptor in the central nervous system, positively coupledwith adenylate cyclase, is antagonized by ICS 205 930. Eur. 1.Pharmacol. 146, 187—188.

Dumuis A., Sebben M., and Bockaert J. (1989) The gastrointestinalprokinetic benzamide derivatives are agonists at the non-classi-cal 5-HT receptor (5-HT4) positively coupled to adenylate cy-clase in neurons. Naunyn Schmiedebergs Arch. Pharmacol. 340,403—410.

Eglen R. M., Wong E. H. F., Dumuis A., and Bockaert J. (1995)Central 5-HT4 receptors. Trends Pharmacol. Sci. 16, 391—398.

Ford A. P. and Clarke D. E. (1994) The 5-HT4 receptor. Med. Res.Rev. 13, 633—662.

Gale J. D., Grossmann C. J., Darton J., Bunce K. T., WhiteheadJ. W. T., Knight J., Parkhouse T. J., Oxford A. W., and Hum-phrey P. P. A. (1994) GR125487: a selective and high affinity5-HT4 receptor antagonist. (Abstr.) Br. J. Pharmacol. 113, 120.

Gaster L. M. and Sanger G. J. (1994) SB 204070: 5-HT receptorantagonists and their potential therapeutic utility. Drugs Future19, 1109—1121.

Gerald C., Adham N., Kao H.-T., Olsen M. A., Laz T. M., SchechterL. E., Bard 1. A., Vaysse P. J.-J., Hartig P. R., Branchek T. A.,and Weinshank R.L. (1995) The 5-HT4 receptor: molecularcloning andpharmacological characterization of two splice vari-ants. EMBO J. 14, 2806—2815.

Grossman C. J., Kilpatrick G. J., and Bunce K. T. (1993) Develop-ment of a radioligand binding assay for 5-HT4 receptors inguinea-pig and rat brain. Br. J. Pharmacol. 109, 618—624.

Huylebroeck D., Maertens G., VerhoeyenM., Lopez C., RaeymakersA., Mm Jou W., and Piers W. (1988) High-level transient ex-pression of influenza virus proteins from a series of SV4O lateand early replacement vectors. Gene 66, 163—181.

Kaumann A. J., Lynham J. A., and Brown A. M. (1996) Comparisonof the densities of 5-HT4 receptors, beta 1- and beta 2-adreno-ceptors in human atrium: functional implications. NaunynSchmiedebergs Arch. Pharmacol. 353, 592—595.

Leung E., Palido-Rios M. T., Bonhaus D. W., Perkins L. A., ZeitungK. D., Hsu S. A. V., Clark R. D., Wong E. H. F., and EglenR. M. (1996) Comparison of 5-HT4 receptors in guinea pigcolon and rat oesophagus: effects of novel agonists and antago-nists. Naunyn Schmiedebergs Arch. Pharmacol. 354, 145—156.

Maniatis T., Fritsch E. F., and Sambrook J. (1982) Molecular Clon-ing: A Laboratory Manual. Cold Spring Harbor, New York.

Oestreicher E. G. and Pinto G. F. (1987) A microcomputer programfor fitting enzyme inhibition equations. Comput. Biol. Med. 17,53—68.

J. Neurochem., Vol. 69, No. 5, 1997

CLONING OF A HUMAN 5-HT4 RECEPTOR eDNA 1819

Sanger 0. J. and Gaster L. M. (1994) 5-HT4 receptor antagonists.Exp. Opin. Ther. Patents 4, 323—334.

Schmuck K., UlImer C., Kalkman H. 0., Probst A., and LObbert H.(1996) Activation of meningeal 5-HT2B receptors: an earlystep in the generation of migraine headache? Fur. J. Neurosci.8, 959—967.

Schoemaker R. G., Du X. Y., Willem A. B., Bos E., and SaxenaP. R. (1993) 5-Hydroxytryptamine stimulates human isolatedatrium but not ventricle. Eur. J. Pharmacol. 230, 103—105.

Schuurkes J. A. J., Van Nueten J. M., van Dade P. J. H., ReyntjensA. J., and Janssen P. A. J. (1985) Motor stimulating propertiesof cisapride on isolated gastrointestinal preparations of theguinea pig. J. Pharmacol. Exp. Ther. 234, 775—784.

Shapiro M. B. and Senapathy P. (1987) RNA splice junctions of

different classes of eukaryotes: sequence statistics and func-tional implications in gene expression. Nucleic Acids Res. 15,7155—7 174.

Ullmer C., Schmuck K., Kalkman H. 0., and LObbert H. (1995)Expression of serotonin receptor mRNAs in blood vessels.FEBS Lett. 370, 2 15—221.

Waeber C., Sebben M., Grossman C., Javoy-Agid F., Bockaert J.,and Dumuis A. (1993) [

3H]-GR113808 labels 5-HT4 recep-

tors in the human and guinea pig brain. Neuroreport 4, 1239—1242.

Wardle K. A., Ellis E. S., Baxter J. S., Kenneth J. A., Gaster L. M.,and Sanger G. J. (1994) The effects of SB 204070, a highlypotent and selective 5-HT4 recepfor antagonist, on guinea-pigdistal colon. Br. J. Pharmacol. 112, 789—794.

.J. Neurochem., Vol. 69, No. 5, 1997